ML20127A968

From kanterella
Jump to navigation Jump to search
Informs of Impact That Risk Evaluation Group Personnel Involvement in Review of Gessar PRA for External Events Will Have on Milestones of Other Branch Projects
ML20127A968
Person / Time
Site: 05000447
Issue date: 10/17/1983
From: Papazoglou I
BROOKHAVEN NATIONAL LABORATORY
To: Thadani A
Office of Nuclear Reactor Regulation
Shared Package
ML20127A367 List:
References
CON-FIN-A-3740, FOIA-85-199 NUDOCS 8310210017
Download: ML20127A968 (2)
v • d • e


Text

{{#Wiki_filter:a.w A a;'L.=.w s.:.=.a.J 4-b rb. ; BROOKHAVEN NATIONAL LABORATORY ff ASSOCIATED UNIVERSITIES, INC. Eb C Upton. Long Isicnd. New York 11973 (516)282s Department of Nuclear Energy FTS 666' October 17, 1983 Dr. Ashok C. Thadani Reliability and Risk Assessment Branch Division of Safety Technology Office of Nuclear Reactor Regulation U. S. Nuclear Regulatory Commission Washington, D.C. 20555

Dear Ashok:

The purpose of this letter is to inform you of the impact that the involvement of Risk Evaluation Group (REG) personnel in the review of the GESSAR PRA for " external events" will have on the milestones of other RRAB projects. Presently, we anticipate performing a review of the " external events" GESSAR PRA of the same scope and level of cffort as the one for the Limerick SARA (Project 3, FIN A-3393). The GESSAR review will involve members of the ~ REG for " system analysis", an outside consultant for the seismic hazard and fragility analysis, and members of the Reliability and Physical Analysis Group (headed by Dr. John Boccio) for the fire physical analysis part of the review. A detailed schedule for the GESSAR review has not yet been established. We anticipate, however, that between November 1,1983 and March 30, 1984, the review will require the equivalent of two months of Dr. Kelvin Shiu's time and two months of Dr. Nelson Hanan's time. They were the leading reviewers of the GESSAR " internal events" PRA. I Dr. Shiu is presently the leading reviewer of the Shoreham PRA and Dr. Hanan is one of the key members of the " System Interactions" team. Owing to Dr. Shiu's involvement with the GESSAR external event PRA, the milestones of the Shoreham PRA review (FIN A-3740) will change as follows: FIN A-3740 Milestones Presently Stated in 189 New Task 1: Draft report on the review 3/15/84 4/30/84 Task 2: Q-1 to applicant 10/31/83 11/30/83 Task 3: Final report Two months after Unchanged receiving appropriate NRC comments i M

'... -a.a.u. o- . Dr. Hanan's involvement might affect some of the system interaction milestones but the exact nature of the effect can not be determined prior to establishing a firm schedule for the GESSAR review. An additional factor that might affect the Shoreham PRA review is the potential participation and/or contribution of Dr. Shiu to the Limerick hearings. I would appreciate it if you could inform us about RRAB's plans concerning the nature and timing of BNL involvement to the aforementioned hearings. 3 I look forward to receiving your response to these comments. Best regards, Y_ _ s, ez -- - _? )

1. A. Papazogl au, Group Leader Risk Evaluatio Group 1AP/dm cc:

A. Busiik (NRC) D. Yue NRC o R. Bari R. Hall J. Boccio

~. - - p XAC 5 6 y MEMORANDUM FOR: Albert Schwencer, Chief Licensing Branch #2 Divis. ion of Licensing FROM: Ashok Thadani, Chief Reliabiity and Risk Assessment Branch Division of Safety Technology

SUBJECT:

SHOREHAM FLOODING W. e have completed the task requested in your memorandum to me dated January 30, 1984 on Shoreham Flooding. With the help of our contractor, Brookhaven National Laboratory (BNL), we have reviewed the internal flooding analysis in the Shoreham Probabilistic Risk Assessment (PRA) study 1 and the Shoreham flooding submittal 2 dated December 2, 1982. Long Island Lighting Company (LILCO) found the Shoreham. core vulnerable frequency (see Enclosure 1, p.4, for definition) initiated by flooding to be about 4x10 s/ reactor year. Maintenance-induced flooding contributes l'.5x10 8/ reactor year to this value, and pipe-break induced flooding contributes 2.4x10 8/ reactor year. For the most part, we found the assumptions and methodology used to be reasonable. However, we have used more recent licensee event report (LER) data and a different model in reevaluating the flood initiating frequency. Our model used a Markov process model to determine the frequency of flood precursor events, and time phased event trees to account for the effects of flooding to different levels. We recognize that there are many uncertainties in the analysis, particularly the human error in initiating a flood and in not taking' proper corrective actions during a flood. We have therefore performed an uncertainty analysis using the SAMPLE 3 program. We estimate that the mean value of the core , vulnerable frequency of accidents initiated by flooding in the reactor building /reactoryear.at Shoreham is 2x10 5/ reactor year, and the 95% upper limit is 7.5x10-The core vulnerable frequency due to ma'ntenance-induced flooding has a mean value of 7x10 8/ reactor year, while the corresponding value for pipe-break induced flooding is 1.3x10 5/ reactor year. Our review identified some potential deficiencies in the Shor'eham alarm-response procedures for mitigating a flood. We note that the human error' probability used by BNL assumed good alarm-response procedures. The core-vulnerable frequency may be higher than that estimated unless the procedures I are corrected. /y@ @Mt9Ci Our fiddings are discussed in the enclosures. is our evaluation; ~ is the preliminary BNL report. We expect to receive the final t CPFICE[ .........o oa"". 5URNAM(h ......... o o o o o o u .n n n o o o o n "" u. i"". D A 78 ) ......."."o' o ". ". " " ". " " ""j ";" [ ;$: ronu m po w sacu no OFFICIAL RECORD COPY ..wi m-b

, a w ,4w- .-m--.+.. 1-se.e ,w 6.r m.i m. p c.. = Id' Q s [ . i KAR 3 0-igg (,x BNL report in the middle of April, 1984, and we will transmit it to you when 5 we receive it. E. Chow (x24727) of RRA8 has performed thi.s assessment.. %~ i 1 Ashok Thadani, Chief + e Reliability and Risk Assessment Branch 3, Division of Safety Technology ~ y

Enclosures:

As stated w j cc: H. Denton I R. Mattson ^ D. Eisenhut T. Speis ". Coffman F E. Chellfah R. Caruso 4 <4 I I /I yl 9 s 3 ~ '\\ 's 4 .'1 .L. (- s i

  • SEE PREVIOUS CONCURRENCE
L A

en.u g .RRAB.:.QSI.?...... 33AS.;.0.5.T...... B.R6Q,; PSI,.\\,,,,,6 p/,D, PSI,,,, A.Thadani

m., l..E..C..h..o..w....m..l.........

...A. 8..u s..l i k ..................... F.R..o..w.. s.o..m e...... RATE l .3./.......g.@. 1.0...).....h../..... 2. 9... /. 8. 4... 3./.... 2 9...../. 6. 4... . 3../.........../. 8. 4... 1 u, . :. f,is s oo. :shacu ana .O F F ICI A L_._R ECO R D _C O P_Y. ess e.....:..

. - _.z_ ,a.c w I,. 'With this assessment, we have completed the task requested in. the memorandum dated January 30, 1984 from A. Schwencer to A. Thadani on Shareham flooding. E.\\C ow-(x24727) of RRAB has performed this assessment. Frank H. Rowsome, Assistant Director for Technology Division of Safety Technology

Enclosures:

.As stated cc: H.'Denton R. Mattson D. Eisenhut T. Speis A. Schwencer . Distribution F. Coffman Central file E. Chelliah RRAB Rdg R. Caruso Echow Chow CHRON ABus11k AThadani FRowsome AD/T Rdg DST CHRON DST 84-22 i 4 \\ \\s r \\ / G,,,c c p...R..R. A..B..: D..S.T....... ..R.R.A..B..:.D..S..T. 5. ..R..R..A..B...:.D..S..T.. ..A..D../..T.. :..D..S..T........................... sunum >...E..c..ho.y Tni ABus11k AThadani FRowsome core >..3../.. 2..'L....[.84. 3 /... 2..'.1...../. 84 ..../..g....'./. 84 ..../............./. 8 4 3 3 a: roau m ommacu cao. OFFICIAL RECORD COPY . w: m-n.

? ENCLOSURE 1 EVALUATION OF SHOREHAM FLOODING 1.0 Introduction A memorandum 4 dated November 16, 1982 on our preliminary review of internal flooding at Shoreham reactor building was transmitted from Stephen Hanauer to Darrell Eisenhut. The preliminary review was performed on the draf t 5 report submitted by Future Resources Associates, Inc. (FRA), the consultants for Suffolk County, and on the draft Shoreham PRA submitted by LILCO. The concern that FRA found that the draft Shoreham PRA underestimated the frequency of certain internal flooding acccident sequences by more than a factor of 1000. 3ased on our preliminary review at that time, we believed that flood-accident sequences did not contribute to risk significantly. However, we recommended LILCO to verify the PRA analysis regarding the followi.ng items: (1) the potential for flooding at Elevation 8 of the reactor building (2) the potential for flood-induced reactor scram (3) the probabilities for e&ch accident scenario based on maintenance schedules and procedures for emergency core cooling (ECC) and. reactor core isolation cooling (RCIC) systems. On December 2, 1982, LILCO submitted an analysis performed by its contractor Science Applications Inc. to respond to the FRA concern on Shoreham flooding. On June 24, 1983, LILCO submitted the final report on the Shoreham PRA which included the most up-to-date analysis on flooding. With the help of BNL,'we have reviewed the December 2, 1982 submittal and the final Shoreham PRA'on the flooding issue. Section 2 discusses some aspects of the data used in the analysis - in particular, the initiating event frequencies and operator error probabilities, including a discussion of alarm-response procedures. Section 3 and 4 discuss the methodology and uncertainty analysis. Section 5~gives the summary and conclusions.

2. 0

' Data Used In The Analysis 2.1 Evaluation of Flood-Initiator Event Frequencies There are two types of initiator events that will lead to flooding of the reactor building at Shoreham. Flooding may be initiated either due to not 1 n,e, -

is u isolating a system which is under maintenance or due to a rupture in the system. What follows is a description of each type of initiator event. 2.1.1 Maintenance-Induced Flood LILCO has obtained operating experience based on LERs8 for turbine-driven pumps and motor-driven pumps in ECC and RCIC systems. The LERs covered events up to 1978. We have also obtained operating experience for the pumps; however, the LERs? that we examined covered events up to 1980. Using the more up-to-date data base on LERs, we estimate higher failure rates for the pumps. These failure rates were used to determine maintenance-induced flood event frequencies. 2.1.2 Pipe-Break-Induced Flood To assess the rupture ' frequency qua'ntitatively, LILCO has considered ruptures of pipes, welds, valves, and pump casings. ~ The general approach that LILCO used to calculate the frequency of a 'lood. f initiated by a rupture in an ECC or RCIC system is as follows: (1) LILCO identified the appropriate type and length of piping and number of components in an ECC or RCIC system susceptible to rupture. (2) LILCO used the LER information in NUREG/CR-13638 and the estimates for leakage and rupture rates in WASH-14003 to calculate the rupture rates for various ECC systems. Our review of BWR operating experience on flooding due to ruptures noted that, in April 1978 at Browns Ferry Unit 3, the supply line to the condensate ring ' header, which provides makeup to the high pressure coolant injection (HPCI) and RCIC systems, failed at a welded joint. Th4 weld failure resulted in flooding of the core spray pump room. LILC0 did not include this event in its data base. We note that the weld at Browns Ferry was mainly made of aluminum whereas the welds.in HPCI system at Shoreham were made of stainless steel. However, we have included the Browns Ferry event in estimating the frequency of flood initiated by ruptures. 2.2 Operator Error Probabilities 2.2.1 Types of Operator Errors Operator errors play significant roles in initiation of a flood and in plant recovery during a flood. The different types of operator errors in a flooding scenario at Shoreham are described as follows: -(1) During a maintenance of a ECC or RCIC pump, an operator may disconnect the electric power to equipment and isolation valves by pulling and O e ,=wwg --..,,,6 --e-y---r ,-y,------,,- -r--*v- ---?'

n . tagging the appropriate breakers at motor control centers. A second person is required to verify that tagging has been perfo'rmed properly. If the electric power to an isolation valve is not removed due to operator errors, and a demand to open the valve occurs during the maintenance,'there would be an open path from the water sources to the reactor building. The demand may be an actual demand for the system or may be a manual demand due to an operator inadvertently operating a switch in the control room. (2) During a maintenance of a pump, an operator may inadvertently by manual local operation open an isolation valve and cause a flood in the reactor . building. (3) When a flood in the reactor building is annunciated by alarms in the control room, an operator may fail to notice the light which is on a back panel. (4) When a flood in the reactor building occurs, an operator must promptly identify the source'of flood and isolate it before it reaches the 3'10" level which disables all ECC and RCIC components. The human error probabilities used by LILCO are based on NUREG/CR-12789 2.2.2 Procedures Review We have reviewed the procedures for operators for mitigating a floo.d. We note that there are specific procedures at Shoreham for detecting and isolating leakages from ECC and RCIC systems. However, we note that the Shoreham alarm-response procedures specify only general guidelines for monitoring system parameters to determine the leakage location and for initiating the leak isolation. The procedures fail to include specific requirements in a checklist for operators to systematically check the operation parameters of ECC and RCIC systems. Since there are many system parameter indicators in the control room, the operators may fail to discover the abnormal system parameters. A checklist with specific steps thst should be followed during a flood in the reactor building would be helpful to operators to reduce confusion and to avoid undue delays in operator responses. Regarding maintenance procedures for pulling and tagging breakers and for . verifying such actions, LILCO stated that these procedures were available for maintenance. 3.0 Methodology Review We have used a Markov model to determine the frequencies of maintenance-induced flood initiators due to maintenance on various components in ECC or RCIC systems. In a similar approach, we have also used another Markov model to determine the frequencies of rupture-induced flood initiators during tra.nsients, manual shutdowns, or' Tests. a 4. ~. .-~

a The analyses submitted by LILCO assumed that when flood reaches 3'10", all ECCS and RCIC components would fail. The LILC0 analysis"did not develop the event trees according to the progression of a flood affecting various components at various elevations up to 3'10". We used a time phased approach to expand the flooding event trees submitted by LILCO into four phases. The four phases correspond to different components at different elevations. Based on the flood rates from various systems, times for the floods to reach various elevations were determined. These times correspond to operator response times for different time phases. The time-dependent human error probabilities were'obtained from NUREG/CR-1278 using the operator response times. The human error probabilities were used to requantify.the event trees for various time phases. 4.0 Uncertainty Analyses In view of the large uncertainties in the analysis, we have used a computer program SAMPLE to estimate the core vulnerable frequency initiated by a flood at Shoreham. The parameters varied in the SAMPLE analysis included: (1) Pipe break frequency (2) Probability of failure of all equipment attached to a division given a failure of a protective relay in a motor-control center (3) Probability of failure of a protective relay (4) Human error probabilities: (a) Probability of failing to rack out a breaker during maintenance (b) Probability of failing to notice a flood alarm (c) Probability of failing to isolate a flood Some of the uncertainties not included in the SAMPLE analysis are: (1) There is no common-mode failure between different divisions, and no sensitivity analysi.s was performed to assess the error here. (2) The conditional probabilities of having a manual trip or a MSIV closure during a flood are subjective and are not varied in our analysts. For example, in our analysis of time phase 4, conditional probability of 0.5 is assumed for a MSIV closure. However, the results cannot be non-conservative by more than a factor of 2. (3) Our' analysis assumes that the Shoreham alarm-response procedures are adequate for proper operator action. Based on the SAMPLE calculation, we estimate that the mean value of the core vulnerable frequency

  • due to flooding is 2x10 5/ reactor year, the

~ *The Shoreham PRA defines the core vulnerable state as an end state of the plant in which the reactor core or containment integrity is challenged. Certain operator actions, including operator actions "in extremes" can be used in a core vulnerable state to prevent core melt. The Shoreham PRA finds that the overall frequency of core melt is about 50% of the overall core vulnerable frequency.

-( - ? upper 95%confidencelimitis7.5x105/reactoryear,andthepower5% confidence limit is 2.2x10 7/ reactor year. We note that the mean value of the core vulnerable frequency due to flooding is about'5 times as large as the estimate obtained by LILCO. The discrepancy is mainly due to our use of higher flood initiator-event frequencies and different approaches (Markov models and time phased event trees). 5.0 Summary / Conclusion Ve find that the mean value of the core vulnerable frequency due to reactor building flooding is 2x10 s/ reactor year. The contribution to this value from maintenance-induced flooding is 7x10 s/ reactor year, and from pipe-break-induced flooding is 1.3x10 5/ reactor year. The upper 95% confidence limit on the core vulnerable frequency was 7.5x10.s/ reactor year, and the lower 5% confidence limit was 2.2x10 7/ reactor year. In contrast LILCO found that core vulnerable frequency initiated by flooding is about 4x10 8/ reactor year; the contribution to this value from maintenance- ' induced flooding is 1.5x10 s/ reactor year, and from pipe-break-induced flooding is 2.4x10 s/ reactor year. Our estimates are predicated upon the assumption that w the alarm-response procedures are adequate. However, we identified some potential deficiencies in these procedures and the core vulnerable frequency may be higher than that estimated unless the procedures are corrected. O D 9 W l } 1 1 O e 8 6 - - - - ~

6 6-References 1. SAI-372-83-PA-01, " Final Report - Probabilistic Risk Assessment Shoreham Nuclear Power Station," Science Applications, Incorporated. . June 24, 1983. 2. Letter from J. L. Smi.th (LILCO) to H. R. Denton (NRC), " Evaluation of Internal Flooding, Shoreham Nuclear Power Station Unit 1," December 2, 1982. 3. WASH-1400, " Reactor Safety Study," October 19,75. 4. Memorandum dated November 16, 1982 from S. H. Hanauer to D. G. Eisenhut, "Shoreham PRA - Review of Suffolk County Consultants - Staff's Preliminary Review." . 5. Letter dated September 24, 1982 from R. J. Budnitz (FRA) to H. R. Denton (NRC), " Review and Critique of Previous Probabilistic Accident Assessments for the Shoreham Nuclear Power Station." 6. NUREG/CR-1205, " Data Summaries of Licensee Event Reports of Pumps at U.S. Commercial Nuclear Power Plants: January 1, 1977 to April 30, 1978," U.S. Nuclear Regulatory Commission, January 1980. 7. NUREG/CR-1205, Rev. 1, " Data Summaries of Licensee Event Reports of Pumps at U.S. Commercial Nuclear Power Plants: January 1, 1972 to September 30, 1980," U.S. Nuclear Regulatory Commission, January 1982. 8. HUREG/CR-1363, W. H. Hubble, C. Miller, " Data Summaries of Licensee Event Reports of Valves. at U.S. Commercial Nuclear Power Plants," June 1980. 9. NUREG/CR-1278, A. D. Swain and H. E. Guttmann, " Handbook of Human Reliability Analysis with Emphasis on Nuclear Power Plant Applications," April 1980. O 9 e sw -m

c l A Preliminary Review 'of the Sequences Following a Release of Excessive water in Elevation 8.of the Reactor Building in the Shoreham Nuclear Power Station K. Shiu Y. H. Sun E. Anavim I. A. Papazoglou March 16, 1984 Risk Evaluation Group Department of Nuclear Energy Brookhaven National Laboratory 4

a

1.0 INTRODUCTION

At Shoreham Nuclear Power Station (SNPS) the majority,of safety-related equipment are located in the Reactor Building (RB). The Shoreham Reactor Building is a cylindrical building surrounding the MARK II containment struc-ture. Water leakage from equipment in the reactor builiding will drain to Elevation 8 (the lowest level of the RB) via openings and stairwells since there 'is no structural se.paration between safety systems. A flooding of'the Elevation 8 compartment may disable the ECCS because the ECCS pumps are installed in the Elevation 8 compartment. The SNPS PRA has included the flooding as a common-mode event which may disable the ECCS equipment. The SNPS PRA assumes that a critical flooding depth of 3'-10" from the RB floor will disable all the ECCS equipment. Operator diagnosis and isolation of the flooding before it reaches 3'-10" depth is considered in SNPS PRA. Because of the potentially significant impact, the SNPS's evaluation of the core melt risk due to RB flooding warrants a special review. A field trip to the Shoreham plant has been made by BML personnel for obtaining detail information on the equipment and power control layouts in the RB, especially in the Elevation 8 conpartment. BNL has determined that there are three floo-ding depths.(l'-3", l'-10", and 3'-10") that are critical to the availabilitj of various ECCS equipments. The initiator event trees are thus revised ac-cordi ngly. BNL also identified that the random failure of a equipment protection circuit breaker coinsiding with the RB flood condition may cause the propaga-tion of failures to equipment poaered by separated Motor Control Centers (MCC). This potential common. mode failure event has also-been modeled in BNL eventi trees. Shoreham Plant Procedure Guides relevant to the RB flooding have been re- ~ viewed by BNL.- BNL found that these procedure guides fail to require a sys-tematic check of system parameter indicators in the control room following a RB Flooding Alarm annunciation. This may cause the operator to ignore a . abnormal system paraneter, especially under a multiple alarm situation (such as a turbine trip). e

_..e_. 2 BNL's revised event trees and the preliminary quantitative evaluation of core melt risk due to the RB flooding event are presented in this report. The report is-organized as follows: Section 2 summarizes the SNPS-PRA ap. i proach to the flood sequence identifications and quantification. Section 3 presents the BNL revision both in the methodology and in t'ne quantification. Finally, Seestion 4.0 summarizes the results. 9 I ' I 4 e G 6 W 6 e T 4 I e O 9 e __,s____

. ~ 2.0 SNPS f4ETHODOLOGY AND ANALYSIS 2.1 Overview The SNPS methodology for determining the contribution to the risk of the internal floods can be divided into three steps. 1. Identification of water sources and pathways to Elevation 8 com-pa rtment. 2. Evaluation of operators responses and assessment of likelihood of ar-resting he flood. 3. Evaluation of system responses and identification of the sequences leading to a core vulnerable state given a flood. In the Shoreham PRA approach it was determined that flooding at locations other than Elevation 8 would be bounded by the analysis of flooding at the lowest level of the reactor building Elevation 8, since the flood water will drain and cascade down to that level through stainiells and' openings. All the evaluations of flood are hence focused on equipment at the Eleva' tion 8 level. The volume of water required,to flood the reactor building Elevation 8 compartment, with all equipment and piping installed, is estimated to be 41,600 gallons in SMPS-PRA for each foot' of depth. The following drainage systems are included to receive the initial volume of flood water. - Reactor Building Floor Sumps Reactor Building Equipment Sumps --Reactor Building Porous Concrete Sumps These systens have total sump capacity of 4,650 gallons, and total sump pump capacity of 640 gallons per minute. The potential water sources which may release excessive water in Ele-vation 8 are summarized in Table 2.1.1. For each of these sources, a pathway investigation has been perforned in the SNPS-PRA, to define the 4 I e ... ~

potential for flood at Elevation 8. Table 2.1.2 summadizes the water sources as evaluated in the Shoreham PRA. For each water source the largest possible flow rate has been determined and the time required for the flood to reach the 3'-10" levels in. Elevation 8, have been estinated. These times are also given in Table 2.1.2. These times provide the basis for estimating the probability of successful prevention of flood at the 3'-10" level by operator actions. A. survey ~of all vital equipment 'by Shoreham identified a number of components fo'r the various accident mitigation systems which could potentially be submerged in the event of an internal flood. Based on this information, the critical height of 3'-10" was defined. It was assumed that if flood water exceeds the 3'-10" level, all ECCS equipment would be disabled. Flooding scenarios which are arrested before reaching the 3'-10" level, have been found to contribute negligibly in the core damage frequency. Functional event trees were used in the Shoreham internal flood PRA to model the plant response given an internal flood initiator. The flood initiator frequency was calculated based on two types of internal-flood precursors: online maintenance and rupture of piping, valves or pumps. These precu'rsors frequencies are described in Section 2.2. Given the occurrence of these flood precursors, the progression of events was modeled using initiator event trees. Details of the initiato'r event trees are presented in Section 2.3. Since all the ECCS systems are assumed lost given a 3'-10' flood, the only available means for cooling the core are the feedwater and the condensate pump i nj ection.. The availability of these two systems depends on the state of the MSIVs'and on the ultimate source of the flood (condensate storage tank or suppression pool). Because of these dependences the end states of the initiator event trees were classified into six categories each of which becones the entry condition for the functional event trees. Table 2.1.1 summarizes the information in a matrix form. Each row of the matrix depicts one of the 17 types of internal ~ 9 =m-

flood precursors, the columns represent the six entry con,'ditions to the functional event trees. The six entry conditions can be grouped into manual [ shutdown, turbine. trip and MSIV closure. Two possible entry conditions are considered for each of these three initiators: flooding due to water from the condensite storage tank ~ (CST) and flooding due to water from other sources. Eased on these six entry conditions, six functional event trees were de-vel oped. An example is given in Figur'e 2.1.1. 2.2 SNPS-PRA Ouantification of the Frecuency of Flood Initiators .Two types of flood initiators were considered in the SNPS-PRA. 1. Floods initiated by an accidental loss of isolation (valve opening) while a component in the Elevation-8 area is dismantled for nain-tenance. 2'. Floods initiated by a rupture in the pressurized or the nonpressurized part of the piping. 2.2.1 Maintenance-inouced Flood Initiators The frequency of an initiator of type one was calculated by estinating the frequency of maintenance of various components from operating experience data. The LER data base in Ref. 2 identifies the observed failures from turbine-driven'and motor-driven pump failures. The data used in the SNPS-PRA are sum-marized in Table 2.2.1. There are four failure modes for pumps i.e., leak-t age / rupture, does not start, loss of function, and does not continue to run. The hourly LER failure rates characterize the leakage / rupture failure mode, whil,e demand failure rates consider other failure modes. The following LER rates are found for the four-failure codes-in motor driven and turbine driven standby pumps. b g J e s -o

d Motor Driven Pumos - Leakage / rupture: 6 events /6,777,627 hrs. = 8.9x10-7/hr. - Does not start, loss of function, and does n'ot continue to run: ~ (5+4+6) events (13,644 demands = 1.1x10-3/deinand) SNPS-PRA assumed that these pumps are in standby status until there is a demand. The number of demand used in SNPS-PRA are 12 on the average per year (four scheduled tests plus eight other occurrences). Hence, the maintenance frequency for motor driven standby pumps per year is calculated as (8.9x10-7 failure /hr) ~(24 hr/ day) (365 day /yr) + (1.1x10-3/ demand) (12 demands /yr) = 2.0x10-2 failure / year. Turbine Driven Puno Similiarly, the maintenance frequency for t,urbine driven standby pumps per year is calculated as 0.079 failure / year. ~ There are two ~ motor driven pumps associated with the Core Spray System, ~ four motor driven pumps with the LPCI System, and four motor driven pumps as-sociated with the Service IIater Systen in which the two are linked as a pair to. the RHR Heat Exchanger System. There is only one turbine driven pump as-sociated with HPCI and RCIC Systems. Table 2.2.2 summarizes the SNPS-PRA frequencies associated with major maintenan.ce operations based upon the above evaluation and a conservative estimate of heat exchanger online_maintenanc.e. e e o, s e....

_ a. _._ ~ . 4-2.2.2 Ruoture-Induced-Flood Initiators The frequencies of the initiators caused by loss of syst'em integrity from breaks or ruptures were derived from WASH-1400 failure rates of major.com-ponents involving external leak and external ruptures, based on assumptions made in NUREG/CR-1363 (Reference 3). This information has been summarized in Table 2.2.3. The calculation of each initiator is done by identifying the appropriate . type and. length of piping and number of components susceptible to rupture and summing the estinated yearly rupt~ure rates. As an example, the total number of valves involved in the HPCI discharge system are 3 (2 M0V's and 1 Check Valve) there is no pump involved (Table 2.2.5) and the total length of piping is 76'. Referring to Table 2.2.3, the rupture failure rate for 100' of pipe section is 4.3x10-ll/hr, and for external failure of a valve is 1.3x10-9/hr. The. total le'ngth of pipe in the HPCI Discharge System is es-timated to be176' (Table 2.2.5). (3 valves) (1.3x10-9/hr)+76'/100'(4.;3x10-ll/hr) = 3.9x10-9/hr or 3.5x10-5/yr. Since t'he flow rates through suction line breaks are time dependent (i.e., a function of the varying water head in'the source) and a strong function of the break shape and size, a simplified model based on historical experience and engineering judgement is used in the Shoreham PRA to describe the con-ditional. probability of break size. Table 2.2.4 sur.marizes the classes of break size examined. These probabilities, are combined with the frequencies estimated for initiators associated with core spray, HPCI, RCIC, LPCI, and Service Water Rupture / Leak Suction System failure to obtain the initiating event frequencies ~ for non-pressurized piping. Table 2.2.6 summarizes the frequencies of initiators due to the loss of system integrity from breaks or ruptures. 9 k ,..mmm,

/g.c4m,4.3 e.- m 4, -ee ud,...- e en weg e au4-p m+ wr mpm. ei.-.#-m-e,-- e >=4

==F w a 2.3 Initiator Event Trees The probability of losing the isolation of a component under naintenance and following that, the probability of not' arresting the flood is calculated with the help of initi'ator Event Trees. These trees are shown in Figures 2.-3.1 through 2.3.17. A' discussion of the P, D, E I, and A events in the ev-ent trees follows. a. Event P - Operator removes power from equipment and valves. The removal of power from equipment and its isolation valves is a re-quired procedure during a maintenance in both fossil and nuclear powe.r stations. The equipment and isolation valves are electrically discon-nected from their associated power supply by pulling and tagging the appropriate breaker at the MCC. A second qualified person verifies the correct inplementation'of the tagging order and placement of the clearance tags. A human error probability (HEP) of 0.01 is assigned for this operator action. This value is determined using the probability data given in NUREG/CR-1278(p.20-23). . b.. Event D - System not demanded. During the maintenance process there is a possibility that the safety systems will be demanded because of a transient challenge. Isolation valves will automatically open if the operator has failed to. remove power from the isolation valves (Event P). c.- Event E - Operator maintains isolation. During on-line maintenance with the equipment disassembled, the isola-tion valves need to be maintained in closed position throughout the duration of the maintenance process. However, an operator error could inadvertently open isolation valves. SNPS concludes that it is unlikely that the operator will manually open these valves locally in the RB and fail to notice the flood. Opening of t.he isolation valves at the MCC is also concluded by SNPS to be unlikely. 9 8 4 4-. +~. .s-g - m,v p-e e

e*-- .- e s.-m ..is%. ,,,,mm, F The remaining possibility is that the valve is opened from the control room (given event P). The panel switch could be activated by three event.s. These events are: the operator mistakenly operates the switch; a command fault to the valve; or the operator inadvertently operates the switch. The probabilities for these events are 10 3, 10-4, and 10-2, respectively. d. Event ! - Flood annunciation. The excessive water in reactor building is annunciated by alarms in the control room. The probability of the operator to fail to notice the alarn (the light is in a "back" panel) is assessed at 10-3, e. Event A - Operator diagnoses and responds to isolate the flood. The operator must identify the source of and isolate the flood before it reaches the 3'-10"*1evel. This event is considered by SNPS under two conditions as follows. 1. Operator isolates flood after auto occurrence, e.g., turbine trip or MSIV closure (Event A ). Multiple alarms will., occur in the A control room at the same time 'as the flood alarm. 2. Operator isolates flood after manual occurrence, e.g., power oper-ation or manual shutdown (Event A ). Only the flood related g alarms will annunciate in the control room. The HEP data provided in NUREG/CR-1278 (1982 Edition, Chapter 12) are' applied by SNPS for their evaluation. Figure 2.3.18 and Table 2.3.1 show the time varying cumulative HEP for both the single and the ~ ~ multiple occurrence conditions. 9 D

1.. _.

Table 2.1.1 Summary of Potential Water Sources'and Types of Initiators Which may Lead to Release of Excessive Water in the Elevation 8 Compartment No. of Source Ouantity (Gallons) Lines Systems -Involved -Suppression Pool 160,000* 8 CS,LPCI,RCI',HPCI C Condensate Storage Tank (CST) 550,000 4 CS,HPCI,RCIC Reactor Primary System ** a) 42,928 b) 152,928 Screerwell (Long Island Sound) Unlimited 4 Service Water Water Fire Protection System. Storage Tank 600,000 Many Fire Main ~

  • Total water volume in the suppression pool at the high water, level mark is 608,500 gallons. However, only a portion of the water'can be drained through ECCS pump suction piping.
    • Figure (a) includes water from the bottom of the core to normal wa'ter level in the RPV. Figure (b) includes (a) plus condenser hotwell water.

4 4 e e e r

Table 2.1.2 Summary of Internal Flooding Initia~tior Types: Source, Pathway, Flowrates, and Time to Critical Flooding Depth ~ Elevation 8 Flooding Time Flow Rate (Minutes *) Source Location gpm* 3'-10" Suppression Pool HPCI Pump Suction 9600 17.6 RCIC Pump Suction 1500 10.6 LPCI Pump Suction (Max /La rge )** 17000/8500 9.4/19.0 CS Pump Suction 13000 12.0 LPCI Pump Suction 10500 15.0 (1 Pump Runout) CS Pump Discharge 6850 23.0 (1 Pump Runout) Condensate Storage Tank (CST) HPCI Pump Suction (Max /Large)** 1200/6000 13.0/27.0 RCIC Pump Suction 2100 76.0 CS Pump Suction (Max /Large)** 1200/6000 13.0/27.0 HPCI Pump Discharge 4350 37.0 (Design) - Service Water. RHR Heat Exchanger 8000 20.0 (PumpRunout) WFPS Rupture of 8" Pipe 4000 ~40.0 ~

  • These flood times were calculated based on a failure of the sump pumps to successfully operate and a 41,600 gallon per foot depth in the reactor building given in the Shoreham FSAR.
    • Large flow rates ~ assumed to be 1/2 maximum flow.

D e .~..y,...

e Table 2.1.3 Sumary of System Event Tree Entry States by Initiator Type SYSTIM [V[NT TRIE !!i1RY L0f.'Offl0N FR(QUCNCY (Per Rs'Tr) IN'ITI AICR H0 M-C T0 T - f. 50 5-C T " 1.0 a 10* " 1.8 10 8 7.6x10 4.3:10'8 ggg T 5.7alo'# 5.7110*I 2.5a10'I 5.0 10 FL2 T 3.0a10 l.1x10 gg3 T 5.0 10*I 4.3a10-6 gg, f 3.6m10 8 6.1x10-8 fL5 1.3x10 T 1.0 10 FL6 3.5s10'I T'g y 6.4al0*I "T 1.la10-5 2.0x10-5' 9.0x10*0 ILS 2.7a10'I 5.8a10*I i 1.3 10 6 gt, T 2.3:10 2.8s10 1.4x10 TL10 T 1.8x10 3.4x10 1.5m10~8 gggg f 1.0x10*I 2.Is10'I fL12 T 2.6s10 7.8a10-8 ~ ggg3 T 1.6 10 2.0a10 8 ~ 8 ftl4 T 4.410*8 2.5s10 ggg$ T 1.lato-6 8.1x10'I 6.6ml0*I itl6 T,Lgg i.4a104 d 8.8:10 2.8 10 10TAL5 1.6a10 5 8.2a10'I 2.2s10-5 3,4,3o 9 1.7a10 5 5.5:10 I . ~,... -

. IU ._ [.. ~ _.._2. Table 2.2.1 LER Data for BWR Standby Pumps for'the Period of January 1972 Through April 1978* Does Not Standby Standby Leakage Does flot loss of Co'ntinue Pumps Demands Hours Rupture Start Function To Run Motor Driven 13,644 6,777,627 6 5 4 6 Turbine . Driven 1,820 868,033 1 6 5 Table 2.2.2 Frequency of Online Major Maintenance-System.in the Reactor Building ~ Frequency (Per Initiator System Year) SNSP-PRA Event Tree Core Spray (Motor Driven) 0.042 TFL3 LPCI (tiotor Driven) 0.'084 TFL4 HPCI.(Turbine Driven) 0.079 TFL2 RCIC (Turbine Driven) 0.079 TFL1 Service llater (RHR or l RBCLW HX) (Motor Driven) 0.042 TFL5 f ~ ~.

Table 2.2.3 Summary of Failure Rates for Major. Components Involving External Leak and External Rupture Total Failur Rupture

  • Parameter Rate Rate /Hr (Mean)

. Reference Failure Rate /Hr . Pipe Failure Section (100') 8.5E-10 WASH-1400 4.3E-11 External Failure of a Valve 2.7E-8 WASH-1400 1.3E-9 External Failure of a Pump 3.0E-9 WASH-1400 1.5E-10

  • Based upon the operating experience to date, given that a failure occurs, the ratio of external leaks to complete failures appears to be in the range.of 20 to 1.

This is substantiated by the specific data review cited in the text for values (18 to 1) and data published by Bush (G-14) on pipes (4 to 1 up to 30 to 1). Because the internal flood evaluation is based upon initiators with substantial flooding rates, i.e., short operator response times, only the catastrophic or large external rupture failures are treated in this ~ . evaluation. Table 2.2.4 Conditional Probability of Pipe Break Size Break Conditional Size Chiracterization Flow Rate Probability Maximum Guillotine Break 100% 0.05 Large Substantial Rupture 50% '0.10 Small* Localized Rupture in Ductile Material 13% 0.85

  • Remainder of the conditional probability was allocated to small breaks.

e. Table 2.2.5 Initiating Event Frequency Estimates. Involving Component Leak / Ruptures ~ VALVES PIPING ~ ESTIM TED INITIATOR SOURCE LENGTH (FT)/ FREQUENCY /, MOV MAN CHK PUMPS SECT /DIA (IN) YR j .neci Disenarge CST /SUPP..2-0 .1 ~' 0 ' ' 76/1/14 3.5E-5 T FL6 CS Discharge SUPP 4 0 2 0 128/2/12 6.9E-5 Tpt7 LPCI Disenarse SUPP 14 4 4 0 240/6/16 2.5E-4 ~'

  • Tygg Service Service Water,

Water 4 4 4 0 715/8/10-20 1.4E-4 'Tygg WFPS. WFPS 1 157/2/6-8 1.1E-5 TFL10 RCIC** 1 1 1 1 70/1/6 3.5E-5 Suction CST T FL11 HPCI** Suction CST ** 1 1 1 1 87/1/15 3.5E-5 IFL1'2,Tptg3 C S ** Suction CST

  • 2 2

2 120/2/12 4.9E-5 Tpt34,TFL15 LPCI** Suction SUPP 4 4 120/2/20 5.2E-5 TFL16,Tpt 7, -CST is assumed to be the source. Suction failures are also classified by-flow rate. e

^ ^ Table 2.2.6. Calculated Frequencies for Initiating Events Resulting from System Ruptures (SHPS-PRA) Initiator Frequency (Per RX Yr) Pressurized Piping HPCI Discharge Break, TFL6 3.5x10-5 CS Discharge Break, TFL7 6.9x10-5 LPCI Discharge Break, TFL8 2.5x10-4 SW Discharge Break, TFL9 1.4x10-4 WFPS Discharge Break, TFL10 1.1x10-5 Non-Pressurized Piping RCIC Suction Failure, TF11 (max) 1.75x10-6. HPCI Suction Failure, TF12 (max) 1.7 5x10-6* HPCI Suction Failure, TF13 (large) 3.5x10-6* CS Suction Failure, TF14 (max) 2.5x10-6. CS Suction Failure, TF15 (large) ' 4.9x10-6. LPCI Suction Failure, TF16 (max) 2.6x10-6, LPCI Suction Failure, TF17 (large)

5. 2x10-3*
  • Modified based upon engineering judgement made on the size of low pressure suction line breaks.

e e ,,e a

- ~. s Table 2.3.1 THE PROBABILITY THAT FLOOD REMAINS UNISOLATED FOR X MINUTES AFTER AUTOMATIC PLANT ACTION: E.G., TURBINE TRIP CR MSIY CLOSURE X P(A (XH P(A M A g 1-1 1.0. 10' 1st + 2nd = 0.54 0.1 20 0.11 0.01 30 0.011 l'.1E-3 60 0.0011 2.0E-4 l 1500 1.1E-4 1.1E-4 ~ ap t e 1

a i ninAem Enu= Ei m aial on.i== caMais saacria - taalt!=e w M u u nw= = tu te. u Au m 8, M,, H tv m uWt 9m m L 58 S Ill8Wella RCIC tiPCI Bt IMS gg ggg ,giv cre'u 5(naes E, ',", at AvAllAttI %vAll AM r ,,E g," I N5 I 8 "' atIt ett e AVAllAelE AVAllAsli 3lggg M 3 (P4 f vtg M e Att I g VAHAMi MW et(Ort et 5 g,,,y g gg.,g, g, s., , C n P e m' u" q a v'- v.

v. "

u. g v. et ( tilpuAll4 - C }' es i .e n,u-t 5.et.) e at., tI A* 5 II 1s384 5; i (1)uUeW*erWusi (t) U U *U* RCIC RA (3) 5.v ev ev"' ~ t e et WI NA 3 ( Itactuate su tuountre isovt) s.e

  • C5 nA 3 l

[ C l.e* i.e =A 1 (

i. -

~ a a , I N*# cl AD 18 1 tehntNSAlt g.y.g 7,gg.go 8.e' l.#c 7 (IA % 11 I,egwway w. I.4t-30 " *'5 a An in i.,,,,,s ii ,,,,.3, f 7 at-5 ',,5,,, e.: n.e " 5 8"

i. t.y inn oncimi M fJ ICIM i g e.g i.-

..__g,,,,, "  !' ' a" _. _.._ n n i= l ,, ( =... I ___._ I ' e. - e..... n e. f Figure 2.1.1 System Event Tree for Manual Shutdowns With Greater Than 3'-10" of Water in the Reactor Building (Source = CST). l k 9

...., _. _.... ~. ..... ~ .~m.. - 8 e g e . 3 0 U C PC .U a e T g e u .e a e J.

====.. = =..... a u E %g C e s e e U. U. U. U. O. O. O. C. U. s.e L r O. z e za =*c a . w. -3 5 a T-e o .o gy- .. =.. 2 7 =. 2

  • M. M.
  • M 32 m

M. 2 w, au -- - ~ ~ w _g 8 I %..2

  • ~

.3 2 W. U U U U m a. h. . g-2 w ,s gggg r .w. rr. gg--- e ,g y av c c -o 5-R Ez$ 13 o 3 = gau As o a. t m e e o o y _= a "E ~ 30..- T = = C --egg e c .hg-4 4 3-4' i5: E B 2 C = = E s ~ e { .!. = .t a m ~~ 3 a r 1 8 8 E. 4 w e o a e C ' f, g E I C.

I_.

C 6 g EE -

=

o =, I 5

  • ~

E gS$ g 6 4 8 = *:g,. a s g B E,E I.-

v. E c

e, s a_ u 3 2 v m2 IaE: g g LIJ - g w g g y g L-r.: a = C 5 -w

a. A e

w A 3 ",st; d 7 3 a o w y

  1. a z

mI$ $ g* O -E r g e !! = = g = g a x =. 5 g a = EY. m. W NU xs ai s. .l 3 - n. 'N

b.
  • 3 W

8

n-y ,3 ~ l: r ) -J s. [ 5 x o r i s) e ..-. z =.zz ... z a= x. = g s I w ky u .e e, .. y. u ...y y. v v eg a za a e z a . A e g g o s-e -- = so

y

= =. 3 .M +. M t + = K =, Er. M. M. ..

  • W t.

W .g 3 ,.m. t o w w

  • ~ -~

- ~ U w ,-. e

= oo c

3

  • E c o y-5

\\ .~,. < , a w~ w w .~ .E 5 '5_E..k 2 hh !E 2 hk 2 3-u. w g g c :r -i e m u ri +r ,e a = a a m.:. = - c = c. u E. L c. w r ~5 g c 5 m-1 5: .e E a g e, s. > = s=52 ~ ~ c 3 3 es a =5V 8 8 E c u O a a e o a U o W. > b E. u u W g= 5 e L -2 w c e w

  1. 3 w

e m wg g e. a .a 3 a 4 -= c E,3 E - a = g fj0g2 - c5 E o ~a M.=- w wg - e as 2 55 0 e = t -x<- = ~ N I 3EIC x i. m 'ul.s le 3* 4 a. j -T -..=m.. = e o c 31 w I e-E L . o E<g E C .e = L g g 0 I. Of We 2 g s = m e .I

... ~, m. a h*. -

===a.... e.e-+-4> m.~,,, pw,_%., g, g-e 4 e e a 9 e-e M E c U e e e e sa s s e sa U .cU U U C-C C to gw M C M

o. e.

a. o. Q e a e e e

o. e.

g ..g g E e4 E W e4 W We& .C .m gg

== -s C E a e c

    • 5 e e

-E ee - - w e - L sp

s. A i. s.
s. s.

.c, e sw. e. 3 .e e. ee a Y 4 g a na C E t w a wg O h - = I-L ,3 5, , a, w5 a =, e-E m m m. a

  • =-

m a a. s s w w w w w "3 oo e o-- o-- C C. U k N b C I: 3* C U I=h w !=G 5:g e. o es w I a' 15 O .C em O Q.- k 6 5 m o ) = I#32 O = a w = -.5<g 15 =a

3. l:ss
f. ;;

% & X " e. s. e b e3-O' C n o

> 6

- '.s uJ L E .; i g .b y g:::, .C .C 5 - m.w a. s. m x. GEG e o a - e C Y I GEE - o = 33* w' I ec -y g E at_ ma O % C j !e 7 .*-]- >~ ~ E. 5 t = g E 5' O - e 88 e 1 - E = "G 8 e. n o 3 ~ga. s I eats = g. M. , gI a g 5-E'. c w a ->=- r s.. 3FE* = Ow b a st"(g s , X. a e a .4 C. 3. y E .e.1, I T U b =r WE

== g = a w .a. a.i g Du g qg .ft - g W E. a.em S*

  • r 3

.m e5

e. -

5 I ' ] S s 9m h e ,-,,n

x m .... s s ss e s a o vcv= Iw c" C g eo oo o sp = z.;. xa ae to 25= m c-4' c v C =

  • *a e

T c a m ~ es - e e e c

  • =
s. a a. a.
  • ?

C .U. sg E ~ t t. 44 44 4 L c .,n W" q ~5 c llE a g5 . *o e o m L a s wS s 0 LEE EE LE E = *: - e a w -e e 5e c -- -- c-u

  • E d-a C

E L (E-43 E 5 'm 6 0 l 5 'g C a f w e g =,n,e, e. o. 4. -2 J g e m g g ,e W 8% =a 6 L .? H = e =

  • G5*a 5

C E EO t g e,' 5 c =. b E TE" c C g ...... > 6 C, 52-LJ 6 3 s - * :=: ua s ~ 5 5 O C =SE.Cag 8, 3 a m m x .a .= E Eg g .e z <5 3 e 5EO = Q 5 &1 E. E. a = 8 = y T .".J a s Ec=1 "E N a-L.= ar H =w-- e e .t.

==*e E Ice E = a t-s r se e 4 - T. m y "a Ee g e* Q E E EE s J

  • C N

I9E*- .a ~' -{ E E * $5 si vi ~ c s a 6 x .2 = E y m ,i .=. Es

m-u s

s U E. i -I "I s. a "f I e. E 9 0

4 .__c__.. e z .. e... o. o. o. e e .t. o R.. m y U G m v v c c o a. 8w - c = w o w: ? ? E a 28 x

  • z e-aa m

e. p C 9 C E s e,. *ll* EW" ? ? TT T ? T b w w w . w w C C $9 55 9 C 9 w _ "-

  • =2 Li.

E w - T L 8 8 0 C w8 gg, "a *a.

  • * ~

=* a e e E -c e~ a w -. $I 8 558h A AA 8I A e-- .E = u-a U g M U C

  • C.

W*Iw I u g a n s. o a 1 E a,. e c C M

  • r 4

3, = O U C E 2},=EwM U *** b b g Mg E "* 'L 3 0 og D_.< 5 c-g a m =tm- _ i

  • c b b l

4 e ai a C o E 8 > b o E= w u = 0 9 : E =s W c 6 .= e 5gw5 5 g. a 4 x 5 = C C U E ~tw Dza e o 2* 4 6J a C a l G E 5-m t c== u -1 ", 3 ~". ..E. al 303 w g SE - O mA c.

    • z.

{ J-I gc=EE i u g E w. I EIET 8 I H 3 ~555 o' o. -I s- . gs WI m.

  • T-R'$

E+ n I a s = s s:x. 4 ~ N 2 4 -.U. i 23_53,. ! e- =3 = t. g=

=

1 : 'I i = o U 32-m 5 y s. = .C. =-

==

u. ~a,

13 = s a .E. m3 as . -.h O 3 E 3 2 YY i I e i i 1 i s l o i k " ~- 4_. ,,y-,n.3,,-%4q

i i l .= 7 i. 1NiflAIOR (VtMT ISEE INIIIAIOR StAtt04 SIAIUS (CON 0lfl0NAL PA00.) BIACIG4 Stillo!NG lhttf.Allt gygg dNt At tW(tgis nR( Al( Of(I IDE(ARICCHH5 fl000 Sif#lENCE CAtttN Aff8 IFPt OF Sf5ftM CM0HIM MIRAIM A5 elmit 5M IM.10 M i Rf,A,C,I,O,R rer = vri Ol5,(tl ARC [io. 0. iiAD.."IS'I II I"* A A H51y AND IIOfe us it i. Al- = iAir. i50AuS io. O. s IO A M^h'38I BNL AIL 51susennst ilmettrE lett (105 tat attorJallt e l$ l I I I A I l L 6' Di* A I I l.It.) F IMA 2.7E 3 M-C N J MAlplAl 5 M ll10W4 (M) l S.I Irma 1.2E 8 5C H ini ,. x.. n-C n m, Ittin 3.2t-8 5-C H s 3.5E.5/ Ot* _E Ur [G"Ott A IllROIN( INIP (I) I 2 ilA e I-C ll "I lit a T-C H 0.17 OK = (CS) l0.01 A' M51V (105ttpt 15A 6.5(-8

  • 5-C H

4.003 1.0 - 158 8.8(-8

  • 5C H

~ OfMAINS Al POWER s0 NtGL IGISLE SIM uaf: M C

  • 8.0t.F. 5-C = 1.X-7 f, '
  • lactu. led la the previewsly evaluated event tree, Figure 2.3.6 lnitiator Event Tree for Postulated Flooding Sequences Initiated by a llPCI Discliarge Pipe Break f

'h. n 3

4 s 4S } m, 9 4"-' ' -emeh%we - _., ~ ' h 0 4 9 E w N S

  • M M

ed Me $ M .E $ M M Wt es o e Q v % M. t - = o. o. o. o. e. o. C eo e s e se.eD E I4 E oft to D's 44 44 Q S De.=. .am. s-m"e a s,un m (,M"

m. =. =.

a. e. e, e = er,n 0 tad t=* w ha e w e a tea 4.a e c= Q e,8= g ed. WS ** en N am ese e gEg W N

  • h N

wy C t,, w C w 3 -.24 % m x5 m a 2 o w5 s "c c*= h; *' 1" 1 x x w - - w.. - .= o = = = = = o== = c== = a a C U - O.

==

ga ac s. n M E SJ E E oo g S;

c. m a

a d 4 = = g g am b f"J a C= g g g* = = 6 U = .52 = + M c

    • 5

.c e-3 =s. = Cp e* G e"* af 3 5 am b

== e >= M C gj _ a

=

v 5 E= a& t., e 4

=

m w mmD* 4 y

IggM a

e m a E 8 8 8 m "" Wg - O O O m e e 2 LT CC g. 3 i c8E t., 3 g @.'O s aa ki e + w y Wt o a wa O U* e E ~w +a 3 o w E aJ '*l 3 w d. g so w as w E.. ~ ",. m h. 3A 9 3 g-m

    • =

= .n. E e., e E a.n. E tas s=b 9 'x"' 3 W 5 I D. w =a=

  • =

= es e e g "u' ' tid e.e S g g h e 4O g O. =.3gT$ lw se =- wt ei .s e= 3 E m E.. = 53E 7 i cm i w = 1 k35~3$$_ E Y C.

t = 3 E

o 4 ,= a ~ g,c+een 2 i a, .m a x w 6 El tea ",i 1 E '- 3 E- _i l i s 6 e, t aEI

==e4 ** e *

  • m

$p e. g -o M 8'" iE 3 ~1 O e e t 8 o e w ~+,e,,,,,.,,-, - ~ = -. -.-,. < t

e e 0 4 m 9 ,E, g e - a a a e a a e a a e we ti, e.s , oy ge e o o o o e o o e o o e

  • t i A i a a?

4 a ox M LO t'" D C1 6 w=g CO e,r, e en. e. e.m m. en. o.m e. s. "O C a O&6 2 3 M. g g4 gg. C= C r et a .= w 99 en N.

  • e w.

w w .L.P e F e L. m =* C wC g ,. O %=E$ 5 C k W3 a - - * * -. e e 3 e abJ m

3. ;
  • 5 5 o - - -.= o w

= w - =- c -g ,c.= oy y O e, to

== 4/'s O' Ee C -= - c. - X *. = 9 v %, r..e. = a"

  • o a

w W 5.me e. = g m a en. g/9 e, Q c. a. s=, m 4 4 4 U + E m" "2 *T s C.= = e e b + g T b em p =m. 8" G~. N* C C o 3 g 3 a.a ~

5... a O

C m u g-< e U~ 3"C9: e > -d = -es ag g - 5 LC "EE C 3- 's w.- e o e o. a. wa

= C . EN. 5 O O ".> 3 o. y M C C E E o a s

  • == U g c.e ac g

f as er - 8 .n -g { OeO*U' O OE g C = a "EE

== E 2.: a. c-~ C" hc2 = g = = = w 3 m =. " 7 =. 2 co. .e.us. r E. n. o-g e .5 S h y.n o $ g .a-- gg en = A NI - d W Q f E3:3 2 E 6 g = ~7 -1 z I. 5 = g l g g., r = = m . = s 2 I 6 z 9 3 3 $ w *W E -i e = 4 s.- ..,%e== M-5 w g go g 3 a g a 6

c~ _ _. z 5 e o e o o e o o e o o, + 3y w o e. o. o. e. o. o. o. va e gy E

  • E
  • e e
m. g C. 6 C. 22 T

.m. >. c -w. ~. ~. ~. ~. ~. e. ~ C: %. llf. 3 *

5. :.

E. :::. w a"s. ,c 5= e e -,- ~ ~ e yyg u w-L oa E E' Ua 22 e a a G "m3*

  • 5

=5 5*t 5 5: 30 0 5 a-s n ~* =U E k aV m-c>

c. L w -
e O

WE a Le 9 q C =

  • s 5!

b 4 w

  • ~

-=s ob i= a o-na = = 'b

= g

1c' = e 2 ag E -03: U .y =. l g-M -4 = e,_ e e o si I' Er o DJ **- >a 1 W $~1g: u.= r :srat E E s o "aag' ;s e' 4 s E ? .a m N U e.

    • ' V

..* $ = n .m., t., a ac E e e-E. go z+5 ~~. + w-e c., = .U. . + - c o o = cw A = ,Io 3 g Z _, a s

r

=w 7 3 a i f 7 m. I = w a M. 1 w-n-.E e ra m a = z y .gs5,= = m a r. = = Si g.* r m. 5 -....o.- o I g = e T L 3lCE{2 55 e = w js 3 5 E b .!! I 5... a"15. g s J e. a=egi a e. ! E i ;y e =* = A l E. 't e ~t a o. ' fl. 3 at 3 .I.- - w =z.y s en Hi!32-1 i 1 = r- -t = e l 8

n e ( .r.e e e eo o e o o e o o. w BW =

e. o.

a. e. e. .g - x z e -e

  • C a

.w. C .e.. .e. e .e .o C.M Y. e e - eg 5** w a. M. E.

3.
  • 3:

M.

  • . =.

Lo g .8 M Pe G P 4 M w.a. - .TC 4 a. g

  • M

.5 m e. = We I "4 3 5 5-er -e .c

-

~ w .c .r -5. wg e-w 2 2 .z. - w .z. w e - bb c. - g - e.= C G O - I .c L' f % k. CT E W. = %o 5 h .~ oe a

  • p

= a e m ce g ws E. e 5-k .h. . -.a. E - a.C = -g 5 = hqf*3a 2-c

E E CM E e a o yo 'wv uC's C = g r.E C e LC r C= g s s:h.= O O -Q= 8 8-

8. g"
  • c a., 4 I

e.e

  • .* M e

a e 's e g a-en e w a 5.,. J. w g E 3 .e -c .e e.. - c. 4 e s 1 g.eP." E. 8' Cg e,. m .E. =G. p y c .e >- s, . o. 3 w a. g .a g E Y a E 3, = h ,,.c.,., n. g Eme"T b h Q g g"

-.= N u e s: 1

= c el gen r.:I 5 = 0 L si s!* I = A g = r CN E **E.2 d p $$E e." $ s " a <we..e.,..-- J. M. w 5E --as fg e. . sw .E, .4 -<3 5 g ge 5 O* '4'* $ a a o E 4, g E a 1 E 3 = 1 = w n.- e. 5 a a E e k .e

5 e i., I i. j i ~ s 5 thill410R (Willl TMt I IstIIAton htAtlam status linuseligouAt rene.) se M ien mesentess Intet.mIlv '*letAt 944 At Smf AE 5494 0ert (AtfeeAltB litt of 515ttM g,,,, .tlC gear,g ges, tu t t*% A5 88fles As f(CHh% Dis ( w w Ge(nMau II't 94 tst? I tt tlHilst? Sitte le(E ' 'l *'l l Ap% Its'I788'5I II8-

10. M 818'M I

15'N All5 kl Allots (Per he Tel t le t ist Ar e t48 ht%64 lb EI584 IS 188 ANeusef l, Alt i. elk 3,, gi gan; 5I All85 pa g gggg A HAggg4g lgg. A Itet. A DESIV jsnetism:re _ e ttet talP flDilmf Nt (fW. Hilt 3 I g , 1,,gg n 1 5 l A R ma l l 1 IMA l.it.le. M-C 8 R Lef-4 i 3 miniAl Semenwnm (n) l*** Inen ~ 5-C R I IMI l.M 9 e M-C l.R e.mi i 183 inte 1.it no 5-C l I. R 4 1 Ilt-6/II To a l l I itmettet f air (1) I# h#ll TIA 9.lt.10 ' 1-C R I g f a cai 8 I 118 2.5t.9 1-C R O.% I M 5[,9,, !" "'I 15A 1.9(-10 ' 5-C 1 R j ,51

5....

5.C l.- L sinums Av reina-a enr.tisistr 1 sittamaf: M C '* l.af-9. T-C = 3.4(-4. 5-C

  • 1.5t.9 1
  • lacle. led la tiv psevleissly evaluated event trees.

t 4 i i 1 i Figitre 2.3.111 Initiator Event Tree for Postulated Flooding Sequences i, Initiatcal l>y a'Haximimi RCIC Suction I.ine Break I i-i I eg 4

a --~w. t 's .= e e v 4 e .Ee .~e s 3 3 3 3 e 3 3 e E 2 e en U1 GJ U w IU o .W e . w w w w w w w w e . e e

e. 5 E e E we

= s= M e c-m- g -e M.hd CU E, W ".. e e. e. c. e I.

9...

U. N. U. U. N. e = 4 W e. O.5., 6 e .c =

  • Co E

u - w m. w.s. a. c= ~ 3 5 u,. I* 2 S5 t eC Ijr E o n .a e - - o - - o - - - z w - - = * * -e = w = a eg - v

== te a, m v -= o-W5 =

c. u-ze e.

c. S e L. e .=. o a sa = .=. y .E a-- e _e a a,a c as c a. Lx 3-M E =s g-1 Hm e e e. m E ri - = c to o o -.a.X { *s* wb .y,ai 3 3 E. I M - e o e I - 3 I E 2 6 "" ec 7 EE E. E " "m I i .e.- .A e mg w o e M * *"* I"'3 3 m ,. C j eJ e w w a.s = 0. I CC -I =.

=

g3 y .as =-

5. w,. < s.

= 4 N O E

  • z* -

I 855a.. r o. u E w n. e O m g 5 .A .i e 28a = 9* 3: G E N e = e.*.$ e GEES 3 s!-s E. I o g e b g W h"15.c i y s W M s.. a i a 3 "1 e., c o 3 l 1 e w a 1 g y =yE I pe 5 w. 2-T g g =- kw y1 I.-

  • a-

= l l 9 e e

  • ~ '~ ~. ~ * ~ ~

w e ,n ,-,~u

e. w< tj. 1 i e I' INillAIGE IWlhl lef t I INillAlas

  1. l ACIM %IAIUS (Cornil1086At P900.)

9t ACl0A sollt Diser. INilcelly arci niin o Wr est As crciers nel Aa nccuns finne stepsutt (Alan tilo ivet er sisiin ,f Sterl tose 45 mItroerA As alsress*J egst 10. Out (ONhiling isrlAAlne RfAtt04 '*'K $8 'I" C l ih[ lt). M il All) 10, ce WL-et5aE I5 IN. 40sNula(I All D 15Hl A,,g $ lt SIAlg$ .sntsK 10 A esumeAs. un l5 IN. A A stilf Aron gg g ) lapCg MaHtsmas IsmalNL I Alt (19511Ni .et(01.Nlif 9 T,g,3 -n 7 s I A a 1 =* In4 2.0(-S M-C H y l NAsma ses>Irans: (nl inAn s.rt-s 5-C u i I Int 4.lf-s M-C H S.003 g 3"I j IttlR 2.6E-s S-C ll I l.St-6/s.v. t OK* I j!E8]N( IRIP Ill IIA s I-C H 0.003 III s I-C 18 OK (5) l 8 Il M5tv Ctestsit( i ISA E.5(-8 5-C 18 I l.o e.Oni 158 1.0C-3 5-C Il j ntleAIN5 AT Poute se ,g CL ygg g, g 5ttetARY! M C

  • 2.6( 5. $-C
  • 7.9E-S
  • 1msluded,lEIN pseviewsly evaluated event trees.

i e I i Figitre 2.3.13 Initiator Event Tree for Postulated Flooding Seqilences Initiated by a large llPCI Suction Line lireak j q-l. 1 i l j t

~-,e.-.. ....__..x. o e B ( e Eg .a .A .a .e a w w a .8 a

  • e>e M

C U.3C E9 I "g

e. o.

e. o. O. O. O. O. Oo e ze z e e e = g ,5 C =". 6 EM Cmc a C E

  • k,
  • O - e C-e

=W e s. m c ,J O y.w =** me e w. e e. X

e was 3-E _E E JJ J d co c-w -

-a L. u w 5_ s M w ":" m =_ C e ~ C s_Ir=. s=_s =_g s___ 5 -x =; -6 3 c. Mm E e

T cc a

g y; q 9

c. 6

= o o o LU E M C ~~~ ~~~ n. sa E Ew UE-E _En 3 h.l =

  • 3_.

c .g ux z;h_- 4 es =< M . a &a ~ g 3UU g E4 C s E E' 5 : s ", 8 -e, S. wo a >h 15 w= w {%. e a 6 g gg a a. ey 55 45 G A &w e wng G ee ~ v5 ~ y i-E, .a_ e s y gw E E *Um. >3 o ~ ~* .C C g {#eN~ 9 E Id -r a e CI M 2 I E = ws e 1.i.s w.;M E. E

  • w=._

S 9 8 T. M . fs 9 8, e y m. - = a e g .,,e, a t I c.y #,-.,,, 5 4 e I g., E w ~ w EEE{5 5 3'. c CE2a E L-55u-:= 5 y a = g = y m <I e.e=-*5 = g e. o 3 ie g E! at s .= 4 g" g w k O i 5fE g a = w g 9 I -y y1 3 =, =- .e., _ i f i l e L

_,.. ~. 0 x na m e e oe o e e e c gg 'i 5 ih ? ? h U *J t e, C 5-

s 2

. a. ?. C' "m'" -n s O EW. e. e. e. e. e. S. Mw EI.

  • W X WW ' * *
  • W M*

mo 6 att A *

  • 4 J J C =-

"O.J C .5 s C= ME f. 3 -C 6-o; e.' 5 - -l E E - ea 6 5~ 5b-3CE 'g y .g .e 5t f "M 4 h Em C_ m -m y-a6 f 9 9 e& "5 ** c e CM e, .I %I o E ]g e. L6 - ::r CC + 2=g -g V

  • U

= e = - E_ 4 ec ce r a m. 66 g g=*x Wm d ', 5 ~ 3:e ~{ 8 8 3 ~ ~ e X :{ 4 e s' y 5, s = s E I sr - w.c m s $55 LT g .,E.s w y Cc 3 [ ua e.g a., to a 5 en = = [.Iw as-g*- g ?, aw ev I ;t '. * ~- 3 at 2 ee e s Er =y W 4 ~= ~~ 32o5e-M g Isi!! I a ~ j a m E7 j a. -kagg a = u y m z E E. m., y','s s L s3 E .-e e J c l 3 e e* = *g f I h ~ a e e 3 =E t z ci m s a w y,, 2 = 5, I-E s. I_ 1 e e 9 e e

s. f. s 4 f W 4 ImIllAlna tving IME 5 InfilAlca al A(Its SIAlWS (fneellienes tene.) M A(50R kNietler. INIfcallt i t(I SuCIIse<nelAEEd(t%*, N4t AE P((tMS '$(AE 9((48* lilW49 llelRAIOR R$A(Ina Ilut talas As kl'.rls:5I 45 M514ptM est lo. On Clemilless l'Al All5 STA1WS Sf Qld llc ( (Al(IEAlle IVPt GF SYST(R sea tsees ie tuellAn5 so.la st-185e415 lu. Assase(I Allh filWS M11&stA10ll Ikilp4N(f Mill 411(( ,F so a essmens '.sels Iw A A mly . p((06 mill > (Per to Tr) Saasteenal latest'al full (105tml T -i 1 ft16 i I M i 5 I A R I SIl6 e 1 OE* letA l.11 6 M0 t j l.0

l. 5.

. i ) 1 0.> t ,,,A, ..M., 50 i 8.003 Sett,

3. P(- 9 M0 L

, 8.3 IMit 5.4(.9 50 L F.68 6thaTe OE* TWR$1NE 181P (1) I yg4

a. Il -7 1-0 L

til F.48 9 1-0 L 8 378 DE g40 ISA 1. 78 7 5-0 L i 1 (1851st( (11 8 II 153 5.0(.10 5-0 L j', M9441815 Al 1cuta s a MGL IGiti t f .ieseAars M-e *

1. ll-6. T e
  • 8. II-F. 5.e. 6.68 -7
  • Intladed in lle reeviently evalveled event trees.

i' l 4 6

  • l' l

Figure 2.3.16 Initiator Event. Tree for Postillateel Flooding Seqlsences Initiated by a Maxiiman 1.PCI Suction Line Ilreak 1 h D*

e.; -t-1 i .L I 184l14I41A (Wth! IM E ^ taillAlemt St ACIce SIAlu$ (feemptileseAt rimes.) M M let RI614DImr. IntfGalIV i t. i rs. l $ser i geniaus As es test', not At arts:RtasM AE K(tutt flefles elet DAlsel M MIsla .l. I184 888 AE 1% pt'artifs51 A5 M5PGII54 Inst 50. et (mentilagt t 85ftl All5 SIAltf5 5(flufffff CAtConAlth f ert of ' 5Y11tel IAmCE e4 ein 4i Ain tit en ht $tt i M?td.fi III Assinger.l A gt i, llette M5l(Jentest tatsyg ge(Y St apellett E0 A leoneAt IN A IHNelage % frelV 4 pitnGMiltl. [per as Tr) I 6 SHlillWEat IIIP (I OSidls( I f T,,,, It I I A R j g } M* I, i I. e.1 tem s.w.y se e " ' 5""' """" I"I !"3 l t samn v.g. s_e l l a e

  1. :810 3 Illi 6.IIF.9 80 0 '

L Ia> i 1. s.. t ? s.PF.6/teTr GE* , l 8 5' 6 lanslut lair (I) ( o.asi lit a,g.g i.e l L M 9851W I Cl M (5) I 154 3.ut: 7 5O L ISI l.00-9 5-0 t

.cct istan t.

. j """'"5"'"'"'"" I t SasseAnt; et8 = 2.4(.F.f.4 = s.g(.F.5-0

  • 2.M-1
  • Incimard en eG*prevleesty evaluated e t tvers.

4 i I I Figure 2.3.17 Initiator Event Tree for Postulated Flooding Sequences Initiated by a 1.arge I.PCI Suction Line Break [

~. ~. _ .__c__..._. g i 1.3

  • ...,,s E

u2,2

  • cKE3-1273 K!Y. 2 F.!?

for single event \\, '\\. - --- NUKEG 1278 KEY.2 HIP for t.e sec:no of :ul:1 i sie events \\ D gg-1 es: \\ c E ..\\ \\ i e.: c '., y'M.TI E EVE'IT io 2 e. = uC I Hm .? O SUICLE EYtMT = at; 1; . ' * ~.. ' M E-to-2 .. g,,g G 10.: 10 20 30 40 50 60 Time (Minutes) . Figure 2.3.18 Cc=parison of the P.EPs Associated with Operator Actions for Singular Events and Coincident Multiple Events e 6 9

p l 3.0 BNL ACCIDENT REVIEW At:0 SE00ENCE QUANTIFICATION This section discusses the quantification and review of the internal floo-ding accident-sequences in-the St PS-PRA due to system maintenance and pipe ruptures. The section is organized as follows. Subsection 3.1 presents a summary of the approach used by BNL to calculate the initiator frequencies. Subsection 3.2 discusses BNL quantitative review of the initiator event trees, and Subsection 3.3 presents the functional event tree analysis and evaluation.

  • ?

4 e S F 9 4 e l I I e L a

o 4 _.__m..__ .-v. 3.1 Flood Precursor Frequency This review revised the assessment of the frequency.'of the flood initia-tors _in two ways. First the experiential data for the estimation of the var-ious failure. rates were revised to include re. cent events. Second, the models for calculating the frequency of floods (or probability per year of reactor operation) have been improved by removing unnecessary conservatisms. As it was already discussed in Section 2.2, two types of initiators were con- ~ sidered: a) maintenance-induced initiators; and b) rupture-induced initiators. The revised frequencies for these types of initiators are presented in the following two subsections. 3.1.1 Maintenance-Induced Flood Initiators A flood can be initiated during the maintenance of a component of the ECCS or of another system in the elevation-8 area, if the maintenance requires dis-mantling of the component and one of the isolation valves opens inadvertently while the component is maintained. The components that contribute to these initiators are the pumps and the heat exchangers in the elevation-8 area. 'These are standby components that can fail in a time-dependent fashion while on standby. Periodic tests are performed to check their operability and if found failed they are put under repair. A Markov model that describes the stochastic behavior of these canponents has been developed and quantified. The important.characterist.ics of this model are as follows:

1) The component can be in six states (see Figure 3.1.1).

i i-) In state 1 the component (pump, heat exchanger) is available, that is ready to start operating if asked to do so. iii) The component while on standby can fail with exponentially dis-tributed times to failure. A failure brings the component into state 2 (see Figure 3.1.1). iv) The failure remains undatectable unt'il a test is perforned or a real challenge is posed to the component. A test.that will find the com-ponent in state 2 will initiate a repair action. The same will hap-pen' following a real demand for the component. b o

v) There are three repair states. States 3 and 3' in which the com-ponent is under repair while the reactor is online, and state 4 where the component is under repair with the reactor sh'utdown. vi) Fol' lowing a test that finds the component failed and before the dis-mantling of the component, all the appropriate motor operated valves musf. be closed and their breakers racked out from the corresponding MCCs. There is, however, a chance that the operator will not recove the breakers from the MCCs leaving then the MOVs able to open fol-lowing a signal to do so. If the probability of such an error is P, then a test brings the component from state 2, to state 3 with probability 1-P (breaker removed) and to state 3' with probability P. vii) The component remains in states 3 or 3' until the repair is completed and then it returns to state 1, or until the allowable outage time is exhausted and then the component transit to state 4 where the repair continues with the reactor shutdown. When the repair is completed, the reactor is brought back online and the component returns to state 1 (transition 4to1). viii) While in state 3', an, actual demand for the component (following a transient initiator) or an inadvertent operation of the corresponding switch in the control room will result into the opening of one of the isolation valves. This event is modeled by a transition of the com-ponet from state 3' to state 5. The reactor transients and the oper-ators errors are assumed to occur with constant rates. A 0 and o, respectively. Quantification of the Markovican model and the determination of the probability that the component will occupy state 5 at the end of one year yields the probability that there will be a maintenance-induced flood by that particular component. 1 l

.. e Quantification The solution of the model requires the quantification of the following parameters.

1) The catastrophic failure rate A.

Th'is' failure mode implies such failures that require major naintenance (dismantling) of the com-ponent. The SNPS-PRA used the data presented in Table 2.2.1 from Ref. 2. BNL has updated this table using additional data included in an updated version of Ref. 2 (Ref. 4). The new data are summarized in Table 3.3.1. Maximum likelihood estinators for the failure rates number of failures A(= - ) yield. total operating time A= 5.7x10-5/hr for Turbine Driven Pumps and ' 4 ~ A= 3.3x10-6/hr for Mot'or Driven Pumps

11) The mean times to repair were assumed 100 hrs and 50 hrs for the turbine driven and the reactor driven pumps, respectively.

Thus U = 10-2/hr for Turbine Driven Pumps and V ='2x10-2/hr for Motor Dri.ven Pumps. iii) In the BNL revision of the SNPS-PRA, the. frequency of transients involvin,g MSIV closure has been assessed at 4.42/yr. Thus, the frequency of transients on an hourly basis is D= 5.0x10-4/hr A iv) Tests are performed every 3 months (4 times a year) for both motor driven and turbine driven pumps. The allowable outage times are 14-and 7 days for turbine driven and motor driven plumps, respectively, y) The probability of not racking out the. breakers of the isolation valves (P) is assessed in the SNPS-PRA as 10-2 The same value is used in these requantifications. vi) The mean time for inadvertently activating a particular switch in the control room has been assumed equal to 10,000 hrs. This implies a rate of A = 10 4/hr. o 8

  • c Quantification of the Markovian model with the numerical values of the parameters mentioned above yield the probabilities per year for the various maintenance induced floods. The results are tabulated in Table 3.1.2.

Additional assumptions are that the. Core Spray System consists of two notar driven pumps, the LPCI consists of four motor driven pumps and that RBCLW heat exchangers are equivalent to motor driven pumps. 3.1.2 Rupture-Induced Flood Initiators A flood can be initiated if a rupture occurs at any point in the pressure boundary of the various systems in the elevation-8 area. Such a rupture will involve one of the following three types of components: 1) piping; 2) valve; and 3) pump. The model assumes that catastrophic ruptures occur in the fol-lowing way. A component fails in such a way that if it is demanded to ope-rate then a catastrophic rupture (large enough to allow the flow rates neces-sary for the flood sizes of " interest to this analysis) will occur. That is,. the component transits first in a rupture-vulnerable state and then, when a de-mand occurs, it ruptures. A Markov model that decribes this stochastic behavior,has been developed and quantified. The model is graphically depicted in F,igure 3.1.2. The basic characteristics of the model are as follows: (1) The system in que.stion (HPCI, RCIC, LPCI, CS, RHR, RBCLWHX) is in state where it is available to perform its function. (ii) The system transits to state 2, which is a rupture vulnerable state A with failure rate R- ~ (iii) If a demand occurs while in state 2 a flood is initiated. A demand occurs whenever a transient, a manual shutdown or a test occurs. We distinguish three flood states: State 3, which is a rupture trig-gered by a transient involving an MSIV closure; State 4, which is a rupture triggered by a turbine-trip transient; and State 5 which is rupture triggered by a manual shutdown or an equipment test. The solution of this model yields the probabilities that the system will occupy states 3, 3 and 5 denc:ed by P, P, Pg, respectively. These S T probabilities at the end of cne-year period provide the frequency of rupture - initiated flood precursors. The expression for these probabilities is 9 .h.

s s A[,A R (1-e-A ")/AR - (1-e-At)j\\ {y) P (t) = F g R where i = S, T, F is the number of tests per year. A j is the rate of arrival of a transient of type 1 (i=S T) A R is the rate of catastrophic rutpure failure in the system and. A is the rate of arrival of any transient ( A =A +A +A ) 3 T M For the manual shutdown the corresponding expression is P (t) = F(- { (1-e-A- )/A - (i.e-At)jy), ,-A T,, T) R M g (2) Ouantification For a given system having piping of length L.n valves n pumps the y p failure rate R is equal to 1 A A A R +"v v*"p p (3) ~ where A, % are the catastrophic rupture failure rates for valves and y pump and ' the same failure rate per unit of piping length. , A search of the LER, has indicated that at least one pipe rupture (weld failure) has occurred in the ECCS piping in the 215 accumulated BWR year. (See Ref. 5). This provides a maximum likelihood estimator for the rupture fai. lure rate 5.31x10-7/lbr). Assuming as in the SNPS PRA that only one of (1/215y = out of every twenty ruptures will create a bre'ak large enough to generate floods of the sizes of concern to this analysis, the catastrcphic piping rup-ture rate becomes A = 2.7x10-8 This of course is applicable for the total length of safety related piping (denoted by L). 8 k- .- - - _ - - - _ - - - - _. _ _ _.._-_____?_

For a particular system with a total of piping lengtih 2, then the catastrophic rupture rate for piping becomes A"=(f)x2.7x10-8/hr (4), where /L denotes the fraction of the total length of the piping that belongs to the particular system. 4 j For the rupture rates of the valves and the pumps, the WASH-1400 values were used (see Table G.4-4 in SNPS-PRA). Using the length of piping, number of valves and pumps provided in Table G.4-5 of the SNPS-PRA, and by virtue of I Eqs. (1) - (3). The total failure rate R for the various systems along with the probabilities P, PT and Pg were calculated. The results'are 3 tabulated in Table 3.1.3. A total of 13.51 transients per year were assumed (4.42 MSIV closures, 4.89 turbine trips and 4.2 manual shutdowns). The splitting between maximum and large floods for i~nitiators TFL12-TFL13, TFL14-TFL15, TFL16-TFL17 was done as in the SNPS-PRA, that is', I to 2. 1 I 9 i l e i 1 o. = .. +. - -

Table 3.1.1 LER Data for BWR Standby Pumps for.the Period of January 1972 Through September 1980 Does Not Standby Standby leakage Does Not loss of Continue Pumps Demands Hours Ruoture Start Function To Run Motor - Driven 13,644 6,777,627 6 5 4 6 Turbine Drivan-1,820 868,033 1 6 5 .,.. d. C.'!' r~'" ~ Table 3.1.2 Frequency of Maintenance - Induced Flood Precursors. i -System Initiator Event Trees Probability per Year-TFL1 P.D 'W A 1.05x10-4 1. RCIC TFL1 P.Ee

s. L '-

2.10x10-5 TFL1 P.E, 2.10x10-5 t TFL2 P.D 1.05x10-4 2. HPLIC -TFL2 P.E 2.10x10-5 o TFL2 P.E, 2.10x10-5 t 3. Core Spray TFL3 P.D 1.89x10-5 (2 motor driven pumps) TFL3 P.E 1.87x13-6 o 4.- LPCI TFL4 P.D 3.78x10-5 (4 mdtor driven) TFL4 P.E 3.74x10-6 n 5. Service Water TFL5 P.D 1.89x10-5 (RHRorRB(LWHX) TFL4 P.E 1.88x10-6 o 2 motor driven pumps e e e - 6 6 s o

2....

Table 3.1.3 Flood Precursor Frequency Pipe Valves Pump Total Ag P3 PT Pg TFL6 1.2(-9) 6.5(-9) 0 7.7(-9) 1.57(-5) 1.7(-5) 1.5(-5) TFL7 2.0(-9) 1.3(-8) 0 1.5(-8) 3.1(-5) 3.4(-5) 2.9.(-5) TFL8 3.7(-9) 2.86(-8) 0 3.2(-8) 6.5(-5) 7.3(-5) 1.2(-5) TFL9 1.1(-8) 2.34(-8) 6.0(-10) 1.29(-8) 2.6(-5) 2.9(-5) 1.5(-5) TFL10 2.4(-9) 1.30(-9) 0 3.7(-9) 7.5 (-6 ) 8.4 (-6) 7.2(-6) TFL11 1.1(-9) 9.10(-9) 1.5(-10) 1.04(-8) 2.1(-5) 2.4(-5) 2.1(-5) TFL12 1.4(-9) 3.90(-9) 1.5(-10) 5.5(-9) 3.7(-6) 4.0(-6) 3.6'-6) 7.3(-6) 8.0(-6) 7.1( 6) TFL13 TFL14 1.9(-9) 5.20(-9) 3.0(-10) 7.4(-9) 5.0(-6) 5.6(-6) 4.8(- 5) 1.0(-5) 1.1(-5) 9.6(-t)- TFL15 TFL16 1.9(-9) 5.20(-9) 6.0(-10) 7.7(-9) 5.2(-6) 5.8(-6)

5. 0 (- 6,'

, 1.2(-5) 1.0(-5) 1.0(-5) TFL17 4 h e 5 r i 6 e t + l 1 9 8 n. ruess,+r

- - ~ ~ 3.2 BNL Ouantitative Review of the Initiator Event Tree The quantitative review of the initiator event trees is discussed in the following subsections. 3.2.1 Review of Flooding Alarm Related Procedures The RB water level is-detected by two RB water level monitors installed on the RB floor. The flood alams are activat.ed by the monitors when the water level is more than 0.5 in, above the floor. The sump alams will be activated ~ when water level reaches the sump alarm setpoints installed at a level right below the level that activates the RB flood alams. Sump alarm sensors are installed at various locations in the RB. .~ The immediate operator action specified in the Alam Response P.rocedure (ARP5671) is to initiate the Suppression Pool. Leakage Return System. The re-quired subsequent actions are: 1. Monitor RB water level to determine approxinate leak rate. Use sump ~ ~ alams to supplement the infomation obtained from the above instruments to ascertain the approximate location of the leak. ~2. Monitor parameters (such as line pressure and flow rate) of the safety systems as a leak would affect the systiem parameters., Isolate the source' of leakage per procedure listed below in 3. 3. If required and plant condition permit,. dispatch an operator to the RB floor to visually locate the source of leakage. Isolate using'th.e ap-propriate system procedure listed below. System HPCI, Procedure No.SP23.202.01 Leakage indication: Abnormal suction or discharge piping pressure. Excessive HPCI Loop Level Pump Flow or low dis-charge pressure. t 1 k e O 9 ' ~ ^

-. ~. -.- = - - u o o Reactor building sump hig,h water levels in vicin-ity of leak. Reactor building flooding alarm. Leakage isolation: If in standby, isolate the HPCI system'by secur-ing the HPCI Loop Level Pump and then closing CST Suction Valve (MOV-031). If the system is operating, secure per shutdown procedure and then isolate as describec above. RCIC, Procedure No.SP23.119.01 Leakage indication: Abnonnal suction or discharge piping pressure. Excessive HPCI Loop Level Pump. Reactor building sump high water levels.

  • Reactor building flooding alarm.

Leakage isolation: If in standby, isolate the RCIC system by secur-ing the RCIC Loop Level Pump and then closing CST Suction Valve (MOV-031). If the system is operating, secure per shutdown . procedure and then isolate as described above. RHR, Procedure No.SP23.121.01 Leakage indication: Heat exchanger service water side temperature inconsistencies. Abnormal RHR system flow for mode of operation. Abnormal RHR system pressures for mode of oper-ation. Reactor water level inconsistencies for mode of ope rati.on. Sump high level alarms'. Reactor building flooding alann. Leakage isolation: Isolate the leakage by shutting down the affected loop in accordance with the apprcpriate procedure e 9 0

for the mode in which it was operating and then systematically shutting valves to isolate areas s of the system fcund above to be possible sources of leakage. s The above isolati.on procedure may require inter-mittent operation of-the leakage return system to observe the effects on water buildup. When the leakaga has been isolated return the un-affected p'ortions (as required) to service. BNL has found that; SNPS alarm response procedures specify general guidelines for monitortNg system par' meters for determining the leakage loca-tion and for initihting th'e leakage isolation. However, the procedures fail to include specific requirements for operators to systematically check the ~ i i operatio1 parameters of relevant systems. Since there are'many system para-neter indicators in the control room, the operator,s r.m possibly fail to ob - serve the indicatio.n of fan abnormal system parebe'ter. i, ; \\' When the abnormai condition is severe enough to actuate the a,larn:;cf a'. /\\6 particular system parameter, the corresponding, Alarm Response Procedure will then be followed by operators) However, BNL has reviewed the relevant Alarm Response Procedures' for abnormal system pa'rameters,'and found tirat Ih'ese procedures do not contain st'eps that should be followed under RB flood con-t ditions. These procedures provide guidelines for conditions other than RB flood, such as water source abnormal or isolation valves abnormal,(etc. The operator responses to the flood could.be. delayed or confused when t,bese Alann Response Procedures.are followed. s 3.2.2 Recuantification l l The revised initiator frequencies are appli,ed for evaluating the sequence frequencies of the initiator event tree. In addition to the critica) flood depth of 3'-10" used by SNPS, BNL also evaluated the sequence frequencies cor-l responding to flood depth of l'-10" ned l'-3". This is because, as indicated I in Table' 3.2.1, flood heights of l'-10" and l'-3" will disable several vital l 5

.e ., A e ~ systems such as HPCI and RCIC. Thetimesforthefloodthreach3'-10", l'-10", and l'-3" depth were calculated based on the leakage. flow rates de-termined in SNPS PRA. The calculated times are shown.in Table 3.2.2.. The HEP. values used by SNPS are identical to the nominal HEP values 'provided in the Probabilistic Risk Analysis Procedure Guide (see F'igure 3.21-and Table 3.2.3). BNL feels that the HEP could be higher than the nominal HEP values because the flooding alarm related procedures fail to provide specific e guidelines to identify and to isolate the flood source (see Section 3.2.1). i The HEPs under the multiple alarm and the single alarm conditions are listed in Tables 3.2.4 and 3.2.5. ~ e 9e 's e o e 4 e e e e 9 e e 9 e 9 9 9

r .- -~. ,9 3.3 BNL Review of Functional Event Tree This section is divided into three subsections. Section 3.3.1 provides a qualitative review of the Shoreham Internal Fl. cod event tree analysis and Sec-tion 3.3.2 presents the BNL revised time phased event trees. Section 3.3.3 describes the results obtained from the quantification of the BNL event trees. 3.3.1 Oualitative Review In general, BNL is of the opinion that the methodology used in the Shoreham Internal Flood Analysi.s is consistent with that of the state-of-the-art and the approach is reasonable. The analysis for the inter-nal flood postulated much severe scenarios than those of the Shoreham FSAR. The Shoreham Internal Flood functional event tree analysis is based predominantly on the event trees developed for the internal event initiators, namely, turbine trip, MSIV closure and manual. shutdown. These internal flood functional event trees only model flood scenarios where the flood water height ~ S at Elevation 8 exceeds 3'-10". While it appears that the Shorehan functional event trees do provide a representative mo'deling of the plant response, it is not wel.1 substantiated that floods that are arrested before reaching 3'-10" will result in negligible core vulnerable frequency. Table 3.3.1 gnumerates the vital equipment that has been identified in the Shoreham analysis. The components are presented with those located at the lowest elevation first. It can be seen that at the l' level, both the RCIC and HPCI vacuum pumps and condensate pumps are expected to b.e disabled. How-ever, it is judged that their failures do not lead to the failure of the re-spective high pressure systems. Similar arguments apply to the loop level pumps of the low pressure core spray, HPCI and the RCIC systems as well. At approximately 2', instrumentation for both high pressure injection systems are submerged an'd hence resulting in failure of both systems. At 3'-10" instrumentation for both LPCS and RHR is submerged leading to the fa'ilure of those low pressure systems. In the Shoreham analysis the critical height of 3'-10" is selected. However, since both HPCI and RCIC have failed at about 2' O 4 e

< L *r ~. 1 level, these scenarios with termination of the flood pr{or to 3'-10" may not contribute an insignificant amount to the core vulnerable frequency. In the BNL revised event trees, a time-phased approach is used to include the con-tribution from flooding below the 3'-10" level. Another area of concern stems from the treatment of propagation of f ailures in the Shoreham analysis. As noted in Table 3.3.1, at th'e l' level, ~4-480V pumps are expected to experience electrical shorts. The Shoreham an-alysis did not investigate any cascading f ailure which may result from the electrical shorts. BNL reviewed the electrical drawings and elementary drawings for some of the systems. It appears that for each pump there is only one electrical breaker which separates it from the rest of the loads in the same motor control center (MCC). Random failure of this breaker to open could result in the propagation of the short circuit fault upstream t'o the MCC, other MCCs and the load center. BNL's review of the electrical diagrams indicates that failure of the breaker to open will res'lt in tripping the u breaker at the load center. Discussions with Shoreham engineers suggested that there may possibly be an additional breaker per pump that is in series with the first breaker. 'However, this was not confirmed-by BNL. In the BNL revised event trees, only one breaker is assumed and its failure is modeled. explicitly. BNL did not review the spraying effects due to water cascades from higher, elevations. O l e e s .y -. -w.-. ,,..--..b-. ,-.---.m _.....,.._.. -, ~.. -.. r

,Ma t 3.3.2 BNL Time Phase Event Tree The determination of the time periods which are critica,1 to the con-sideration of the progression of the flood is based on the vital equipment location list (Table 3.3.1). Three heights were selected for the BNL anal-ysis: at the l'-3" level, at the l'-10" level, and at the 3'-10" level. If the flood is terminated prior to reaching the l'-3" level, no impact is as-sumed for any equipment and the plant will be shutdown, this is Phase 1. How-ever, if the flood water exceeds the l'-3" level, but is terminated before the l'.-10" level, this is Phase II. Phase III entails the failures of both HPCI and RCIC system as well as the loss of power to the MG set recirculation pump fluid coupler before arresting the flood below the 3'-10" level. Any flood level which exceeds the 3'-10" level, it is treated in Phase IV. The event trees of these four phases are presented in Figures 3.3.1 througn 3.3.4 Given that the flood is terminated in Phase I, BNL assumed that the reactor has a high probability (0.9) that it will be manually shut-down. Ten percent of the time, it may result in a MSIV' closure event. These two branches of the Phase I event trees are transferred to the respective internal event tree, Figure 3.3.1. Figure 3.3.2 depicts the Phase II functional event tree, in which the var-ious. mitigation systems are considered. Moreover, owing to the fact that a number of the 480V pumps will be flooded,'the possibility of a breaker failure to isolate the fault.is also evaluated. It is assumed that the breaker fail-ure to open probability is 1x10-3 and there are a total of five pumps in Division 1 and two pumps in Division II that will be short circuited. A prob-ability of 0.5 is also ass'uned that failure of a load center in a division -would lead.to failure of other equipment connected to that division. In the event of a MSIV closure, the feedwater system is considered to be unavailable. The probability that the reactor will be manually shutdown is also assumed to be 0.9 for the maintenance induced flood events. Figure 3.3.3 illustrates the functional event tree used to describe the Phase III events. The major difference between this event tree and the Phase 11 tree is the high pressure systems. In the Phase III events, both the RCIC and the HPCI systens are not unavailable due to the failure of respective instrunentation. The probability that the reactor will be manually shutdown is assumed,to be 0.5 for the.naintenance induced flood events. 7

o%< The Phase IV event tree is presented in Figure 3.3.4. This tree is drastically different from the other ones in that it only considers the feedwater system, the depressurization function and the PCS. All the other systems are disabled due to fl'ooding. The liklihood that the reactor will be ^ manually shutdown is the same as in Phase III for maintenance-induced floods. 3.3.3 Ouantitative Analysis Based on the development of the revised flood initiator frequency, BNL time-phased event tree and the modified human response to arrest flood, prelininary quantitative results are obtained. There are 17 different flood precursors. Similar to the Shoreham classification, the first five precursors are online maintenance related; the remaining twelve of them are rupture re-lated. A detailed discussion on the BNL flood precursors is given in Section 3.1. ~0 wing to the ways that these flood precursors are calculated, the ini-3 tiator event trees have been modified to include only three, functions: the flood alarm annunication, I; operator action to isolate flood, A;. and reactor status. The entry condition to the different time phase event trees is deter-mined by the A function (see Section 3.2 for details). Each of the 17 flood precursors were evaluated with the initiator event tree and the four time phase event trees. The' unavailability values for the various event trees are the same as those used in the Shoreham analysis except' as noted in the last section. } When the time phase e' vent trees were quantified for the 17 flood pre-cur ors, the results are the conditional frequency of core' vulnerable given the particular f,lood precursor. These frequencies are summarized in Table 3.3.2. The seventeen precursors are listed as rows while t'he four phases are shown as columns. Within each precursor, contributions from manual shutdown, MSIV closure or turbine trip are also shown. For instance, the conditional frequency of core vulnerable with operator arresting the flood prior to 3'-10" but after l'-10" - Phase III, for TFL1 is 2.0(-5) given the reactor is man. ally shutdown. liowever, if instead of a manual shutdown, the plant experiences a MSIV closure, then the conditional frequency is 8.5(-4). 4 e

.~.. -.. o *T As expected, the conditional frequency consistently inpreases as the flood progresses to higher el'evations. In other words, the conditional frequency of Phase IV is always larger than any of the other phases. Another noteworthy observation is the unusually large conditional f'requency of core vulnerable for the LPCI system induced flood, i.e.+. TFL4 and TFL8. The TFL9 and TFL5 values are also large since they disabled the LPCI systems as well. The core vulnerable frequency given the BNL revised flood precursors, initiator event trees and time phase event trees is shown in Table 3.3.3. In this table, the 17 precursors are' depicted on the left with the 4 phases de-picted as columns. Each precursor also identifies the contributions from the various plant states. Core vulnerable frequency contributions from Phase I ~ and II are very small, in the order of 10-9.~ Contributions from Phase III are not insignificant but not substantial, approximately 10-6, 3,yenty,per- . cent of the total core vulnerable frequency (70% of 2.0(-5)) is attributabl'e . to LPCI system maintenance or rupture induced flood. The'naintenance con-tribution to flood is about 37% while the balance is due to rupture. It appears also that failure to properly 'model the fault propagation of 'the short circuits through the breakers does not have a significant effect on core vulnerable frequency. 9 O e 8 9 9

^ o by o 4.0

SUMMARY

BNL reviewed the internal flood analysis which is a part of the Shoreham PRA and found that assumptions, methodology, and results are reasonable. ENL revaluated the flood precursor frequency using recent LER data and a more accurate methodology. This methodology avoids some of the conservatisms in the SNPS-PRA approach. A slight increase in the initiator frequency is calculated because of the revised data. .Similarly, based on the PSA Procedure Guide, the HEP was reviewed and only minimal changes were made to the Shoreham HEP values used in the analysis. As for the functional event trees, a time phase approach was adopted to better model the progression of the flood events. Results are summarized in Tabl,e 4.1. This table can be divided into two . pa rt s. Part A provides a comparison between the Shoreham results and those obtained in the BNL review. The BNL value is about 5 times that of the Shoreham frequency, 2.0(-5) vs. 3.9(-6). The contributions from the different plant states are also presented. Part B of Table 4.1 compares only the con-tributions from the BNL Phase IV results with the Shoreham values. It can be inferred that by neglecting the initial three phases, the core vulnerable frequency will be underestimated by 3x10-6,or about 18%. The major increase in core vulnerable frequency in the BNL analysis is attributable to the increase in flood precursor frequencies. e 5 ~

4% o ) 48 4 REFERENCES 1. PRA Shoreham Huclear Power Plant, LILCO, June 24, 1983. 2. NUREG/dR-1205, " Data Summaries of Licensee Event Reporf.s okf Pumps at U. S. Commercial Nuclear Power Plants: January 1,1977 to April 30, 1978," U. S. Nuclear Regulatory Commission, January 1980. 3. W. H. Hubble, C'.11 iller, " Data Summaries of Licensee Event Reports of Valves at U. S. Commercial Nuclear Power Plants," NUREG/CR-1363, Volume 1, June 1980. 4 s* ~? 9 9 9 O I 6 G O G l

_.m-s. ....... = - j- ,f 3 ro va nerva n run a nt nurvira ur :._ _. _-y 1984 q FIN: A-3740 TITLE: Shoreham PRA Review LAB: BNL TECH. MONITOR: E. Chow PERIOD OF PERFORMANCE: Monthly Business Letter Received: Yes /No CONTRACT SCOPE /0BJECTIVE: The objective of this contract is to evaluate the Shoreham PRA study in order to datermine the risk profile at Shoreham in comparison with the risk profile at the referenced BWR in WASH-1400. MAJOR MILESTONES / SLIPPAGES: 9g gD& n /5 /ss i 3 /'s M+ .] w 4 6 /g MQ" l (c 4 kg & +ho n+ k (@ C4 Q & 2 /3- /14 s l 3 1 ?] CURRENT STATUS / PROBLEMS - PROPOSED RESOLUTIONS: &%Q (. e W $ 4'ESSAR

  • Y W N N

&p

  • & cQR W &f c4 U

[ g gs 4 A-tL w; tL sR ee ~i k Y e p-SQ Q j ggAS W b r yA4hh%WM+hE j 6 q tL C Pr# % - l FY84 .i I FY84 DST Budgeted: 10 .g ] FY84 Obligated: 'l) l. Coments:

.f) .. n u. - ~ ;. 3 : s e us xi.eux i r ux i nt. Mun i n ut r1 ARCH 1984 ~, FIN: A-3740 TITLE: Shoreham PRA Review LAB: BNL 2// /* ' ~ a[/ No' TECH. MONITOR: E. Chow PERIOD OF PERFORMANCE: Yes-Monthly Business Letter Received: y }} } CONTRACT-SCOPE /0BJECTIVE: 1

The objective of this contract is to evaluate the.Shoreham PRA study in order to j

determine the risk profile st Shoreham in comparison with the risk profile at the /{ referenced.BWR in WASH-1400. Ij q MAJOR HILESTONES/ SLIPPAGES: N (t\\ Inytd.1 qawi4iw Y*M-m' "w [ (z) C-s%p12D sv% 6 { M'"-r da-lua mv j (3) Com M 1rk-ie7X 44 /VF }3 (M C,qAIL pj t<p.,T sh./w } i il S

s CURRENT STATUS / PROBLEMS - PROPOSED RESOLUTIONS:

.. a $nd L %,mdvwh 7]v 'h - \\]8d ih }. .? J !I%v y /j

?

FY84 9 j FY84 DST Budgeted: 87/s[ FY84 Obligated: / #/)/i Coments:4- / 2.v / ~, 7 / h j u -# ION l .i O

4,' 'j blAlub RtPURI FOR lHL NUNIH Ut bA%), l'; $ i j FIN: 'A-3740 TITLE: Shoreham PRA Review Lt.B : BNL TECH. MONITOR: E. Chow PERIOD OF PERFORMANCE: Monthly Business Letter Received: Yes /No 'tj CONTRACT SCOPE /0BJECTIVE: i The objective of this contract is to evaluate the Shoreham PRA study in order-to d3termine the risk profile at Shoreham in comparison with the risk profile at the referenced BWR in WASH-1400. MAJOR MILESTONES / SLIPPAGES: g (o Pnnd g - 4vJfM M - l ' /5M (2) y % 4 M% - 3 /is /8+ n) c@ 44t g +ho/8+ + yc 4 a /v / 8+- w. 1 lij Il 1 i CURRENT STATUS / PROBLEMS - PROPOSED RESOLUTIONS: 'k I '. l2 g t { c, 6% w M g,y d MM j ] bM Y WC o 19 .4 [I q. ?! FY84 i FY84 DST Budgeted: 2.7/) FY84 Obligated: /MO k Coments : f2't> +m o ElM9/ub/ Aeby<//99bjm2N b .j J a $ ~f/ A 9/ 5*Y a

I' ~2 _. a __. __ a._. j bl AIUS RLPORl FOR lHL MUNIH Uf o n3 FIN: A-3740 TITLE: Shoreham PRA Review 1 LAB: BNL ///.333'4/.3089 ' TECH. MONITOR: E. Chow PERIOD OF PERFORMANCE: Monthly Business Letter Rec ived: Yes No l,* l CONTRACT SCOPE /0BJECTIVE: i The objective of this contract is to evaluate the Shoreham PRA study in order to determine the risk profile at Shoreham in comparison with the risk profile at the q referenced BWR in WASH-1400. i MAJOR, MILESTONES / SLIPPAGES: lt (O Ettn <)4 d e h I e) c,yAt % e.. AJ % - 3 /is A+ f n) cqht <tc{i p +Ao/s+ MC d q.p(- 8 / 50 / 8 & ) g t ii CURRENT STATUS / PROBLEMS - PROPOSED RESOLUT!r?c: 6 N' L i. c d d ah Q } A 6NL M <b.l.3 nM M]N A 9' fl L 3 1

1 q

tl d ~! FYS4 FY84 DST Budgeted: 8N[/ FY84,0bligated: /4///d# Comments: -E h 4, in /L - /?:/ n +- .y c,?,, f /,.o n. w w//,e i ~ ' ~ l- ./<sp d ls'c/ '$2'i/Sl

.. ~ STATUS REPORT FOR THE MONTH OF FIN: A-3740 TITLE: Shoreham PRA Review LAB: BNL TECH. MONITOR: E. Chow PERIOD OF PERFORMANCE: Monthly Business Letter Received: Yes/No CONTRACT SCOPE /0BJECTIVE: The objective of this contract is to evaluate the Shoreham PRA study in order to determine the risk profile at Shoreham in comparison with the risk profile at the referenced BWR in WASH-1400. M d MAJOR MILESTONES / SLIPPAGES: d-M (t) PM a A y I (z) C '.s N 483d4 9 c ' n) c@ @ syd f ho /s9-t 1 CURRENT STATUS / PROBLEMS - PROPOSED R'ESOLUTIONS: GNL 's 34M iY ' g {-f}u % & % Sb YW O n cfG _d . A 4T'A t5dl eM $Y ^ sw *

  • M

%4% p w e c J.Ls m tydy D eGf vw q% t, -t/~ epdkw FY84 FY84 DST Budgeted: _. Pi'/t' } FY84 Obligated: /.c X Coments : 4 L

I STATUS REPORT FOR THE MONTH OF DECEMBER 1983 FIN: A-3740 TITLE: Shoreham PRA Review LAB: BNL TECH. MONITOR: E. Chow PERIOD OF PERFORMANCE: Monthly Business Letter Received: Yes/ o N CONTRACT SCOPE /0BJECTIVE: The objective of this contract is to evaluate the Shoreham PRA study in order to detennine the risk profile at Shoreham in comparison with the risk profile at the referenced BWR in WASH-1400. MAJOR MILESTONES / SLIPPAGES: W U!D S (t) P qs M 'i b (2} h V3in toc + y,x f I CURRENT STATUS / PROBLEMS - PROPOSED RESOLUTIONS: 4 b \\h vMw j BNL 8W eq 7Y) M G4 f%ch BNL w r r d t * >-, FY84 FY84 DST Budgeted: FY84 Obligated: Coments:

STATUS REPORT FOR THE MONTH OF qwg H i3 FIN: A-3740 TITLE: Shoreham PRA Review LAB: BNL TECH. MONITOR: E. Chow PERIOD OF PERFORMANCE: Monthly Business Letter Received: Yes/No CONTRACT SCOPE /0BJECTIVE: The objective of this contract is to evaluate the Shoreham PRA study in order to determine the risk profile at Shoreham in comparison with the risk profile at the referenced BWR in WASH-1400. 4 '\\ i MAJOR MILESTONES / SLIPPAGES: Y (t) P g pv 4 N +8MI4 (z) 9 l (Q & & v78 1 ho /84-i 1 l CURRENT STATUS / PROBLEMS - PROPOSED RESOLUTIONS: Br4L'sqM p' b 54 Y% Y$ W

  • GQ My% 6dL 4 W A m y

i @ ~ 4 4 Ac e t & e -t tL.< A % FY84 FY84 DST Budgeted:

    1. //

i FY84 Obligated: /f// Coments: N

~ - ~.. STATUS REPORT FOR THE MONTH OF O c-6 F!N: d 'J N O TITLE: 5HOREtiAM f'RA ret'iEW' 'I LAB: 3/ll TECH. MONITOR: 6 PERIOD OF PERFORMANCE: Monthly Business Letter Received: Yes / o N ci CONTRACT SCOPE /0BJECTIVE: f $ A f et f & h kEa % tub $ ~kla SS PO h Le k'E M L a & ~+ W w + A -rL M rat 4 pj Mb SWR a v/ ash 1400-m ?) MAJOR MILESTONES / SLIPPAGES:

j (g) fdQ M

I'

  • N m
  • }

a) cgtk. 4 m k +/s h+ i o) gpp e hc /s+ l !), CURRENT STATUS / PROBLEMS - PROPOSED RESOLUTIONS: Q-> g & &Q k GrGSSA tt f'RA A Q SARA, SHL o Mky W fv A

  • 7 f[

[\\ een 8ch - h l FY84 FY84 contract Cost: d36N + l FT84 DST Budgeted: ph) A%m da,u)L/We/ FY84 Obligated: Coments: # 8h 3 -+ l'J & R <url ohE a N,wss n) 5 sw IM $ ' d rf up,,'

> i n l ub t uk I HL I;utilH Ut SEPTEttBER 83 IN: AOY TITLE:

/J#/7 'AB: /3/2)/ ECH. MONITdR: Eb PERIOD OF PERFORMANCE: so/19/f 5 Tc 6/8 /T& Monthly Business Letter Received: Yes .ONTRACT SCOPE /0BJECTIVE: ! 'Y c4.h d Csdd ' M b S !u d n F R A M A 'd C ! % %Me%L~t e xu p A-tLa 4 d Nv S (C w/A f H-MLoo. ii ~jAJOR MILESTONES / SLIPPAGES l'j

0) fd % t $1p CM i M /s+

.j ,] (.N N t/i/ F4 ) ti) c4 u4t44 mrt a,ue+ (+3 c@ p g ,/,s/r+ 1 q ,j URRENT STATilS/ PROBLEMS - PROPOSEU RESOLUTI0iiS: i l: Qs A &HL a sh t & - sp m bo GM h ~}}v & wrh.. Cbbn fRA M A H i- : I l l FY83 FY83 Contract Cost FY83 DST Budgeted 56 / FY83 Oblicated . f/)M i Total contract Cost c745F 4 ?:1 l; F ' es

B 3' [p ato D UNITED STATES o

  • /

g ~g ' QNyCLFAR REGULATORY COMMISSION g,7 + a g g 7, .; y fgg wasnincron, o. c. rosss ( ,f t < J . 'h. \\,f DEC 2 91983 s 0 / (>,,. 'f [ i f 's. ,/ 3o q'*%l%0 g.c2 o 0 > t,. 3: e. m% r+ \\ r h., %y 4 MEMORANDUM FOR: Harold R. Denton, Director / Office of Nuclear Reactor Regulation c FROM: Roger J. Mattson, Director N Division of Systems Integration l

SUBJECT:

SHOREHAM REACTOR BUILDING - INTERNAL FLOOD PROTECTION 4 At the invitation of R. W. Starostecki, Director of the Division of Project A and Resident Programs, Region I, I participated with Region I personnel in 9,j4 an on-site assessment of the adequacy of internal flood protection at the 4 Shoreham plant. I was accompanied on this December 6,1983, site visit by L. S. Rubenstein and E. Sylvester of my staff, who have been involved in the ongoing discussions with Region I. g o The Region's question of flood pr'otection adequacy first came to my attention h[ in a memorandum to D. G. Eisenhut dated June 3,1982, from Mr. Starostecki in c which he requested NRR assistance to resolve several outstanding safety issues at Shoreham. We subsequently provided a safety evaluation report of our. / understanding of this concern to Mr. Starostecki by memorandum from T. Novak (' dated May 9,1983. After several telephone conversations with Mr. Starostecki, N it was decided to meet with him and other Region I personnel to pursue the issue. At the December 6 site visit, we met with Mr. Starostecki, Mr. Charles Petrone, Resident Inspector for Shoreham, and Mr. Thomas Shedlosky, Senior Resident Inspector for Millstone, who has been assisting Mr. Petrone. They identified three separate internal flooding problems: (1) post-LOCA equipment leakage in the reactor building, (2) moderate and high energy pipe break flooding, / k [ and (3) flooding due to procedural errors during maintenance of reactor building fluid system components. After a tour of the facility with Long i Island Lighting Company personnel, we and Region I personnel agreed that the safety-related reactor building sump pump provided adequate protecticn against the minor leakage expected after a LOCA. We (NRR and Region I) also agreed on the adequacy of the protection afforded essential equipment in the reactor building from pipe break flooding. We consider these two aspects of the flooding concern to be resolved. However, we concluded that further evaluation will be required to resolve the concern as it relates to flooding from main-tenance procedure errors. Internal floods resulting from maintenance procedure errors are currently beyond the scope of our deterministic review process. The scenario for Shoreham reactor building flooding postulates maintenance activities whereby fluid systems components are opened to the reactor building atmosphere, an r b ~~~ I

L-3. o Harold R. Denton DEC 2 91983 / operator erroneously opens isolation valve (s) to the component, and there is a failure to teminate the ensuing leak in time to prevent flooding of essential equipment. By memorandum from S. H. Hanauer to D. G. Eisenhut dated November 16, 1982, the staff documented an evaluation of the draft Shoreham probabilistic risk assessment of this accident sequence along with an evaluation of the Suffolk County consultant's report on the Shoreham PRA. The staff concluded that the maintenance flooding sequences do not contribute to risk significant1v, subject to applicant verification of plant-speciric event orobabilities._ The evaluation was sent to tne Atomic safety ana Ticensing Board for Shoreham by November 26, 1982 memorandum from T. Novak. The applicant has subsequently submitted a revised PRA of the potential for flooding due to maintenance errors. This _ submittal, dated December 2,1982_, has not_been eyalua.ted_by__the staff. The Divis1orFof Licensing has agreed to ~ initiate an evalt:ation of the submittal by the Reliability and Risk Assessment Branch of the Division of Safety Technology to ascertain whether it confinns_ 0 the staff's oreliminary conclusion that the maintenance error ~typ_e_of flooding ~ ~ TX not an undue risk. The regional and resident personnel will be kept in- ~ ^ Tomed of the olitcome of that reviqw, projected by DST for conclusion by the end of February,1984. Messrs. Spets, Eisenhut and Starostecki have concurred in this approach. e) Cr budAd (' Roger J. ttson, Director c Division of Systams Integration cc: D. G. Eisenhut T. Spets R. Starostecki L. Rubenstein F. Rowsome T. Novak O. Parr A. Thadani A. Schwencer J. Wilson C. Petrone R. Caruso E. Sylvester r -}}

Retrieved from "https://kanterella.com/w/index.php?title=ML20127A968&oldid=4270723"